CN100517303C - Partitioning ownership of a database among different database servers to control access to the database - Google Patents

Partitioning ownership of a database among different database servers to control access to the database Download PDF

Info

Publication number
CN100517303C
CN100517303C CN 01822844 CN01822844A CN100517303C CN 100517303 C CN100517303 C CN 100517303C CN 01822844 CN01822844 CN 01822844 CN 01822844 A CN01822844 A CN 01822844A CN 100517303 C CN100517303 C CN 100517303C
Authority
CN
Grant status
Grant
Patent type
Prior art keywords
database
partitioning
ownership
control
different
Prior art date
Application number
CN 01822844
Other languages
Chinese (zh)
Other versions
CN1505788A (en )
Inventor
詹弗兰科·普措卢
Original Assignee
甲骨文国际公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Grant date

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/30Information retrieval; Database structures therefor ; File system structures therefor
    • G06F17/30286Information retrieval; Database structures therefor ; File system structures therefor in structured data stores
    • G06F17/30575Replication, distribution or synchronisation of data between databases or within a distributed database; Distributed database system architectures therefor
    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F21/00Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
    • G06F21/60Protecting data
    • G06F21/62Protecting access to data via a platform, e.g. using keys or access control rules
    • G06F21/6209Protecting access to data via a platform, e.g. using keys or access control rules to a single file or object, e.g. in a secure envelope, encrypted and accessed using a key, or with access control rules appended to the object itself
    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F21/00Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
    • G06F21/60Protecting data
    • G06F21/62Protecting access to data via a platform, e.g. using keys or access control rules
    • G06F21/6218Protecting access to data via a platform, e.g. using keys or access control rules to a system of files or objects, e.g. local or distributed file system or database
    • G06F21/6227Protecting access to data via a platform, e.g. using keys or access control rules to a system of files or objects, e.g. local or distributed file system or database where protection concerns the structure of data, e.g. records, types, queries
    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/46Multiprogramming arrangements
    • G06F9/50Allocation of resources, e.g. of the central processing unit [CPU]
    • G06F9/5083Techniques for rebalancing the load in a distributed system
    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F2221/00Indexing scheme relating to security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
    • G06F2221/21Indexing scheme relating to G06F21/00 and subgroups addressing additional information or applications relating to security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
    • G06F2221/2117User registration
    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F2221/00Indexing scheme relating to security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
    • G06F2221/21Indexing scheme relating to G06F21/00 and subgroups addressing additional information or applications relating to security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
    • G06F2221/2141Access rights, e.g. capability lists, access control lists, access tables, access matrices
    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F2221/00Indexing scheme relating to security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
    • G06F2221/21Indexing scheme relating to G06F21/00 and subgroups addressing additional information or applications relating to security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
    • G06F2221/2145Inheriting rights or properties, e.g., propagation of permissions or restrictions within a hierarchy
    • GPHYSICS
    • G06COMPUTING; CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F2221/00Indexing scheme relating to security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
    • G06F2221/21Indexing scheme relating to G06F21/00 and subgroups addressing additional information or applications relating to security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
    • G06F2221/2149Restricted operating environment

Abstract

至少一部分数据库(250)被划分成所有权组(230,232,234,236),其中每一个都被指派了一个或多个数据库服务器(208,210,212)作为所有者。 At least a portion of the database (250) is divided into a set of ownership (230,232,234,236), each of which is assigned one or more database servers (208, 210) as the owner. 指派为所有者(s208,s210,s121)的数据库服务器(208,210,212)被看作是这个所有权组中全部数据项的拥有者。 Assigned to the database server owner (s208, s210, s121) of (208, 210) is seen as the ownership group owner of all data items. 就是说它们被允许在所有权组中直接访问数据项,而其他数据库服务器则不准许直接访问这些数据项。 That they are allowed to access the data item directly in the ownership group, while other database servers are not allowed to directly access these data items. 一个数据库系统中包括一个或多个保存了数据库(250)的永久存储设备(214,216),以及运行在多个可以直接访问永久存储设备(214,216)的节点(202,204,206)上的多个数据库服务器(208,210,212)。 A database system includes one or more permanent storage device holds a database (250) (214, 216), and a plurality of nodes running permanent direct access storage device (214, 216) of (202, 204) multiple database servers on a (208, 210). 至少一部分数据库(250)被划分成了多个所有权组(230,232,234,236),其中每个所有权组都被指派了一个所有者集合。 At least a portion of the database (250) is divided into a number of title sets (230,232,234,236), wherein each group is assigned ownership of the owner of a set. 只有在归属于一个所有者集合的数据库服务器上运行的进程才准许直接访问这个所有权组内部的数据。 Only on the database server belong to the owner of a set of processes running before allowing direct access to the data inside this ownership group.

Description

在不同数据库服务器之间划分一个数据库所有权以控制访问数据库 A database ownership divided among different database servers to control access to database

相关申请 RELATED APPLICATIONS

本申i奮涉及1998年12月28日提交的在先美国申请09/222,577,其名称为"Hybrid Shared Nothing/Shared Disk Database System",发明人是Gianfranco Putzolu,本发明还涉及与本申请同时提交的美国申请09/896,373,其名称为"Partitioning Ownership of a Database Among Different Database Servers to Control Access to the Database", 发明人是Gianfranco Putzolu。 This application relates to i Fen December 28, 1998 filed prior US Application No. 09 / 222,577, entitled "Hybrid Shared Nothing / Shared Disk Database System", inventor Gianfranco Putzolu, the invention also relates filed concurrently herewith U.S. application No. 09 / 896,373, entitled "Partitioning Ownership of a Database Among Different Database Servers to Control Access to the Database", inventor Gianfranco Putzolu.

技术领域 FIELD

本发明涉及数据库系统,尤其涉及一种在不同数据库服务器之间对数据库所有权进行划分以控制访问数据库。 The present invention relates to database systems and more particularly relates to a partitioning of the database ownership between different database servers to control access to the database.

背景技术 Background technique

多处理计算才几系统是那些包括多个处理单元的系统,这些处理单元能够相对于彼此而并行执行指令。 Multiprocessing system is that only a few computing system comprising a plurality of processing units, the processing units with respect to one another and execute instructions in parallel. 为了使用并行处理能力,一个任务的不同方面可以指派给不同的处理单元。 In order to use parallel processing capabilities, different aspects of a task can be assigned to different processing units. 在这里, 一个任务的不同方面称为工作区组(work granule),而负责在可用处理单元之间分配工作区组的进程贝'J称为协调器进程。 Here, the different aspects of the task is called a work area group (work granule), the allocation process is responsible for the workspace shell groups among available processing unit 'J called a coordinator process.

多处理计算才几系统通常分为三种:共享一切的系统(shared everything system )、共享石兹盘的系统(shared disk system )以及无共享的系统(shared nothing system )。 Only a few multi-processing computing systems are generally divided into three types: shared everything system (shared everything system), shared stone hereby disk system (shared disk system) and non-shared system (shared nothing system). 在4寻工作分配给冲丸4亍工作区纽_ 的进程时所放置的约束条件是根据所涉及的多处理系统类型而变化的。 Finding constraints in a 4-work assigned to pellet red right foot workspace 4 _ New process is placed in a multi-processing system of the type involved varies.

在共享一切的系统中,所有处理器上的进程都可以直接访问系统中所有的动态存储设备(下文中通称为"存储器,,)以及所有的静态存储设备(下文中通称为"磁盘,,)。 Shared everything systems, processes on all processors have direct access to the system for all dynamic storage device (hereinafter known as "memory,,) and all the static storage device (hereinafter known as" disk ,,) .

因此在一个共享一切的系统中,对于如4可指派工作区组而言几乎是没有约束条件的。 Therefore, in a shared everything system for such purposes can be assigned to 4 groups workspace almost no constraints. 然而需要在不同计算机组件之间进行高度布线以提供共享一切的功能。 However, the height of the wiring is required between the various computer components in order to provide shared everything functionality. 另外,对共享一切的架构来说,还存在着可伸缩性的限制。 In addition, all of the shared architecture is, there is also the scalability limitations.

在共享磁盘的系统中,处理器和存储器归合成节点。 In shared disk systems, processors and memory node normalized synthesized. 共享磁盘系统的每个节点自身即可构成一个共享一切的系统,其中包含了多个处理器和多个存储器。 Each node shared disk system can itself constitute a shared everything system, which includes multiple processors and multiple memories. 所有处理器上的进程都可以访问该系统的所有磁盘,但是只有属于某个特定节点的处理器上的进程可以直接访问该特定节点的存储器。 Processes on all processors can access all the disks in the system, but the process is only belong to a particular node processor can access the memory of the specific node directly. 共享磁盘系统所需要的布线通常少于共享一切的系统。 Shared disk system needs wiring usually less than all of the shared system. 然而,共享磁盘系统更易受到不均衡工作负栽条件的影响。 However, the shared disk system are more susceptible to uneven workload plant conditions. 举例来说,如果节点具有一个进程,该进程作用于一个需要大量动态存储器的工作区组,那么,属于该节点的存储器可能不足以同时保存所有需要的数据。 For example, if the node has a process that requires a lot of work applied to the dynamic memory block, then the memory may be insufficient to belong to the node while preserving all required data. 因此,即使其他节点保持大量可用 Therefore, even if a large number of other nodes remain available

和未使用存储器,该进程也还是要与节点的本地存储器交换数据。 And unused memory, the process still have to exchange data with local storage node.

共享磁盘系统提供了造成存储器毁损的软件故障划分。 The system provides a shared disk memory damage caused by the failure of software division. 唯一的例外是那些由节点间的锁定管理器所使用的控制块,实际上,这些控制块在所有节点都得到了复制。 The only exceptions are those by the lock manager control blocks between the nodes is used, in fact, all the nodes in the control block have been replicated.

在无共享系统中,所有处理器、存储器以及磁盘都将归合成节点。 In shared nothing systems, all processors, memory, and disk will go synthesis node. 与共享磁盘系统一样,在无共享系统中,每个节点自身都可构成一个共享一切的系统或是一个共享^兹盘的系统。 And shared disk systems, in the absence of a shared system, each node itself can constitute a shared everything system or a ^ hereby disk systems shared. 只有那些运4亍在 Only those shipped right foot in 4

20某个特定节点上的进程可以直接访问该节点内部的存储器和磁盘。 20 process on a particular node can directly access the node memory and internal disk. 在这三种常身见的多处理系统中,无共享系统通常在不同系统组件之间需要最少的布线。 In the three body often see multi-processing system, shared nothing systems typically require a minimum of wiring between the various system components. 然而,无共享系统最易受到不均4軒工作负栽条 However, no sharing system most vulnerable to uneven workload planted Article 4 Hin

件的影响。 The impact member. 举例来"i兌,在某个工作区组中所要访问的全部数据都可保存在某个特定节点的磁盘上。因此只有运行在该节点内部的进程可用于执行该工作区组,即使是在其它节点上的进程都保持空闲的情况下。 For example to "i against all the data in a workspace group to be accessed can be stored on disk for a particular node. Therefore, only the process running in the interior of the node can be used to perform the work area group, even in processes on other nodes in the case remain idle.

无共享系统提供了导致存储器和/或磁盘毁损的软件故障划分。 No sharing systems division provides a software failure results in memory and / or disk damaged. 唯一的例外就是对依据不同节点的数据子集"所有权"加以控制的控制块。 The only exception is controlled on the basis of the control block different nodes of a subset of data "ownership." 与共享磁盘锁定管理信息相比,对所有权很少进行修改。 Compared with the shared disk lock management information, ownership rarely modified. 因此,所有权技术要比共享磁盘锁定管理技术更加简单可靠,因为它们并不具有对于高性能的需要。 Therefore, the ownership of the shared disk than the lock management technology is more simple and reliable, because they do not have the need for high performance.

运行在多处理系统上的数据库通常分为两种:共享磁盘的数据库和无共享数据库。 Run the database on a multi-processing system is usually divided into two types: shared disk database and shared-nothing database. 在共享磁盘的数据库系统中,多个数据库服务器(通常运行在不同节点上)能够读取和写入数据库的任何部分。 In a shared disk database systems, multiple database servers (typically running on different nodes) to read and write to any part of the database. 共享^f兹盘架构中的数据访问是借助一个分布式锁定管理器来协调的。 Shared data access disk architecture ^ f hereby by means of a distributed lock manager to coordinate. 共享磁盘的数据库既可运行在无共享的计算机系统上,也可运行在共享^?兹盘的计算机系统上。 Shared-disk database can run on non-shared computer system can also be run in a shared ^? Hereby disk on a computer system. 要想在一个无共享计算机系统中运行一个共享磁盘的数据库,那么可以向操作系统添加软件支持,也可以提供附加硬件而使得进程能够直接访问远程磁盘。 To a non-shared computer system running a shared disk database, you can add software to the operating system support, and can also provide additional hardware to enable a process to access remote disk directly.

无共享数据库假设:只有当数椐包含在一个与进程同属一个节点的磁盘上的时候,该进程才能直接访问数据。 No shared database hypothesis: only when the number noted on the disk included with a process belong to a node, the process can directly access the data. 具体地说,数据库数据是在可用教:据库月l务器之间细分的。 Specifically, the database data is available to teach: It is between January l library service is broken down. 每个数据库服务器只能直接读写该数据库服务器拥有的那部分数据。 Each database server can directly read and write the data portion of the database server owns. 如杲第一服务器试图访问第二服务器拥有的数据,那么第一数据库服务器必须向第二数据库服务器发送消息,以使第二数据库服务器代表它来执行数据访问。 Gao as the first server tries to access data owned by the second server, the database server must first send a message to the second database server, so that the second database server to perform data access on its behalf. 无共享数净居库既可以运4亍在共享磁盘的多处J里系统上,也可以运行在无共享的多处理系统上。 Shared-library number Jingju either shipped right foot 4 on multiple J in a shared disk system can also run on non-shared multi-processing systems. 为了在共享f兹盘的冲几器上运行一个无共享数据库,可以提供一种软件机制来对数据库进行逻辑划分并将每个分区的所有权指派给某个特定节点。 F hereby order on the shared disk punch several runs a shared-nothing database, the software may provide a mechanism to logically divide and ownership of each partition of the database assigned to a particular node.

无共享系统和共享磁盘的系统各自都具有与其特定架构相关的便利优点。 No sharing system and shared disk systems each have their specific advantages and convenient infrastructure related. 举例来说,如果存在对于数据的频繁写访问(写热点), 那么无共享数据库提供了更好性能。 For example, if you frequently write access to the data (write hot spots) exist, no shared database provides better performance. 如果存在频繁的读访问(读热点),那么共享磁盘的数据库提供了更好的性能。 If there are frequent read access (read hot), then the shared disk database provides better performance. 并且如上所述, 在出现软件故障的情况下,无共享系统提供了更好的容错性。 And as noted above, in the event of a software failure, non-shared system provides better fault tolerance.

根据上文,较为理想的是提供一个单独的数据库系统,该系统能够提供这两种数据库架构的性能优势。 From the above, it is desirable to provide a separate database system that provides performance advantages of both database schema. 然而,这两种架构通常是相互排斥的。 However, both architectures are usually mutually exclusive.

发明内容 SUMMARY

在这里提供了一个数据库系统,其中将数据库或数据库的某个部分划分成所有权组。 In the database there is provided a system in which some portion of the database or into ownership group. 每个所有权组都被指派了一个或多个数据库服务器,使之作为所有权组的所有者。 Each group is assigned ownership of one or more database servers, so that as the owner of the ownership group. 指派为所有权组的所有者的数据库服务器被看作是属于该所有权组的所有数据项的所有者。 Assigned as the owner of the ownership group of the database server is seen as the owner of all data items belonging to the ownership group. 也就是说,这些数据库服务器准许直接访问该所有权组中的数据项, 而其他数据库服务器则不准许直接访问这些数据项。 In other words, these database servers permit direct access to data items in the ownership group, while other database servers are not allowed to directly access these data items.

才艮据本发明的一个方面,提供了一个数据库系统,该系统包括一个或多个持久存储设备,其上保存了一个数据库,该系统还包括运行在多个节点上的多个数据库服务器。 According to a Gen only aspect of the invention, there is provided a database system which comprises one or more persistent storage devices, stored on a database, the system further comprising a plurality of database servers is running on multiple nodes. 每个节点都能够直接访问持久存储设备。 Each node can directly access the persistent storage device. 至少一部分数据库划分为多个所有权组。 At least a portion of the database into multiple ownership group. 每个所有权组都指派了一个所有者集合。 Each ownership group are assigned an owner collection. 只有那些作为所有权组所有者集合 Only those collections as the ownership group owner

中的成员并在数据库服务器上运行的进程才准许直接访问所有权组内部的数据。 The members and on the database server processes running before allowing direct access to internal data ownership group.

22每个所有纟又组都标记为无共享所有4又组或是共享》兹盘所有权组。 All 22 of each group are labeled as Si and all four shared-nothing and shared groups or "Have disk ownership group. 每个无共享所有权组都从数据库服务器中指派了一个所有者。 Each shared-ownership group are assigned an owner from the database server. 只有每个无共享所有^又组的所有者才准许直接访问无共享所有权组内部的数据。 Only owners of each share ^ and no group before allowing direct access to non-shared data inside ownership group. 每个数据库服务器都准许直接访问所有权组内部的数据,而这些所有权组则标记为共享磁盘所有权组。 Each database server are granted direct access to internal data ownership group, and these ownership groups are marked as shared disk ownership group.

附图说明 BRIEF DESCRIPTION

本发明是借助实例来描述的,但这并不作为限制,在附图的图 The present invention is described by way of example, but not by way of limitation, in the accompanying drawings FIG.

形中,相同的附图标记代表相同的部件,其中: Shaped, like reference numerals represent like parts throughout the same, wherein:

图1是可以执行本发明一个实施例的计算机系统的框图; FIG. 1 is a block diagram of one embodiment of a computer system of the present invention may be performed;

图2是才艮据本发明一个实施例而使用所有权组的分布式数据库系纟充的冲匡图; FIG 2 is only an embodiment according to the present invention, Burgundy embodiment using the red-based distributed database Peter FIG ownership of Si charged groups;

图3是一个描述对支持所有权组的系统的^:据项执行操作的步骤的流禾呈图; Figure 3 is a description of the support system of the ownership group ^: Wo data stream performing an operation entry step of FIG form;

图4是一个对根据本发明实施例来改变所有权组的所有者集合的步骤加以描述的流程图;以及 FIG 4 is a flowchart of the embodiment to be described embodiment of the present invention, the step of changing the ownership of the owner of the group set; and

图5是一个描述根据一个本发明实施例来进行原子变化的技术的框图。 FIG 5 is a block diagram depicting a technique embodiments of the present invention varies depending on atomic.

具体实施方式 detailed description

在这里描述了一种方法,用于在不同的数据库服务器之间对一个数据库的所有权进行划分,以便对数据库访问加以控制。 Here it is described a method for partitioning a database of ownership between different database servers, in order to control access to the database. 在以下 In the following

描述中,为了说明的目的,对许多细节进行了阐述,以便提供关于本发明的全面理解。 Description, for purposes of explanation, numerous details are set forth in order to provide a thorough understanding of the present invention. 然而对本领域技术人员来说,本发明显然可以在不具备这些特定细节的情况下实施。 However, the skilled artisan, the present invention will be apparent without practiced without these specific details. 为了避免不必要地使本发明不清楚,其他实例、公知结构和设备是以框图形式显示的。 In order to avoid unnecessarily obscuring the present invention, other instances, well-known structures and devices are shown in block diagram form.

硬件综述 Hardware Overview

图1是描述可以4丸行本发明一个实施例的计算4几系统100的框图。 FIG 1 is a block diagram depicting calculation of Example 4 may be a few system 100 embodiment 4 of the present invention pellet rows. 计算机系统100包括一条总线102或是用于传递信息的其他通信结构,并且包括一个与总线102耦合并用于处理信息的处理器104。 Computer system 100 includes a bus 102 or other communication structure for communicating information, and includes a processor 104 coupled with bus 102 for processing information and. 计算4几系统100还包含一个主存储器106,例如随才几访问存储-器(RAM)或是其它动态存储器,该存储器与总线102耦合,用于保存信息以及处理器104所要执行的指令。 4 several computing system 100 also includes a main memory 106, for example, with only a few access memory (RAM) or other dynamic memory, coupled to the memory bus 102, and command information to be executed by processor 104 for storage. 在运行处理器104所执行指令的过程中,主存储器106还可用于保存临时变量或是其它中间4言息。 In operation the processor 104 executing instructions, a main memory 106 may also be used to store temporary variables or other intermediate information statement 4. 计算才几系统100还包4舌一个只读存4诸器(ROM) 108或其它静态存储设备,其与总线102耦合,用于保存静态信息和涉及处理器104的指令。 Computing system 100 only a few of the tongue 4 a further read-only memory 4 various devices (ROM) 108 or other static storage device coupled to bus 102 for storing static information and instructions for processor 104 relates. 并且提供了诸如磁盘或光盘这种存储设备110, 其与总线102耦合,用于保存信息和指令。 And Providing such as magnetic or optical disk storage device 110 coupled with bus 102 for storing information and instructions.

计算机系统IOO可以经由总线102而与阴极射线管(CRT)这类显示器112相连,从而将信息显示给计算机用户。 The computer system may IOO Such displays are connected via a bus 102 and a cathode ray tube (CRT) 112, thereby displaying information to a computer user. 包括字母数字及其他4安键的输入设备114与总线102相连,以^便将信息和命令选择传递给处理器104。 4, including alphanumeric and other keys An input device 114 coupled to bus 102, to put ^ information and command selections to processor 104. 另一种用户输入设备是光标控制116,例如鼠标、轨迹球或光标方向键,用于将方向信息和命令选择传递给处理器104并且控制显示器112上的光标移动。 Another type of user input device is cursor control 116, such as a mouse, trackball, or cursor direction keys, for communicating direction information and command selections to processor 104 and for controlling cursor movement on display 112. 这种输入设备通常在第一轴(例如x)和第二轴(例如y)这两个轴上具有两个自由度, 由此i殳备能够确定一个平面上的位置。 Such input devices are often (e.g., x) and a second axis (e.g., y) that has two degrees of freedom in two axes of the first shaft, whereby i Shu apparatus capable of determining a position on the plane.

本发明涉及使用计算机系统100来提供一个混合的共享磁盘/ 无共享的数据库系统。 The present invention relates to the use of computer system 100 to provide a mix of shared disk / shared nothing database system. 根据本发明的一个实施例,处理器104执行主存储器106中包含的一个或多个指令的一个或多个序列,计算机系统100对此做出响应,由此提供这种数据库系统。 One or more sequences according to one embodiment of the present invention, the processor 104 executes one or more instructions contained in main memory 106, the computer system 100 in response thereto, thereby providing such a database system. 这些指令可以从诸如存储设备110这样的另一种计算机可读介质读入主存储器 These instructions may be read into main memory from another such computer readable medium such as storage device 110

24106。 24106. 通过执行主存储器106中包含的指令序列,处理器104执行这里描述的处理步骤。 By executing sequences of instructions contained in main memory 106, processor 104 performs process steps described herein. 在替换实施例中,硬布线电路可用于取代软件指令或是与之组合,由此实现本发明。 In alternative embodiments, hard-wired circuitry may be used in place of or in combination with software instructions, thereby achieving the present invention. 因此,本发明的实施例并不局限于硬件电路和软件的任何一种特定组合。 Thus, embodiments of the present invention is not limited to any specific combination of hardware circuitry and software.

这里4吏用的术语"计算机可读介质"是指〗壬何一种参与向处理器104提供指令以供执行的介质。 4 officials herein, the term "computer readable medium" refers to any non〗 a participatory providing instructions to processor 104 for execution medium. 这种介质可以采取很多形式,其中包括但不局限于:非易失介质、易失介质和传输介质。 Such a medium may take many forms, including but not limited to: non-volatile media, volatile media, and transmission media. 举例来说, 非易失介质包括光盘或磁盘,例如存储设备110。 For example, non-volatile media including optical or magnetic disks, such as storage device 110. 易失介质包括动态存储器,例如主存储器106。 Volatile media includes dynamic memory, such as main memory 106. 传输介质包括同轴电缆、铜线和光纤,其中包4舌构成总线102的线。 Transmission media includes coaxial cables, copper wire and fiber optics, wherein a tongue constituting the bag 4-wire bus 102. 传4俞介质还可以采取声波或光波的形式,例如无线电波和红外数据通信中产生的信号。 4 YU transmission media can also take the form of acoustic or light waves, such as radio wave and infrared data signal generated in the communication.

举例来i兌,计算才几可读介质的通用形式包括:软盘、软磁盘、 硬盘、》兹带或任何其它》兹介质、CD-R0M或4壬何其它光学介质、 穿孔卡、纸带、或具有孔洞图案的任何其它物理介质、RAM、 PROM 和EPROM、 FLASH-EPROM、其它任何存储芯片或盒式磁盘机、 如下所述的载波或是计算才几可以读取的其它4壬4可介质。 I against the way of example, calculation of only a few general form readable medium comprising: a floppy disk, a hard disk, "or any other tape hereby" hereby media, CD-R0M, or 4-nonyl any other optical medium, punch cards, paper tape, or other 4-nonyl medium 4 may be any other physical medium, RAM, PROM, and EPROM, FLASH-EPROM with patterns of holes, any other memory chip or cartridge machine, or a carrier wave as computing only a few can be read.

不同形式的计算才几可读介质涉及向处理器104传递一个或多个指令的一个或多个序列,以便加以执行。 Only a few different forms of computing-readable medium directed to pass one or more instructions to processor 104 one or more sequences, in order to be implemented. 举例来说,起初可以在远程计算机的》兹盘上携带指令。 For example, initially be carrying instructions on "Have disk remote computer. 远程计算机可以将指令加载到它的动态存储器中,并且使用调制解调器经由电话线来发送指令。 The remote computer can load the instructions into its dynamic memory and send the instructions over a telephone line using a modem. 计算机系统100本地的调制解调器可以在电话线上接收数据并且使用红外发射机来将数据转换成红外信号。 A modem local to computer system 100 can receive the data on the telephone line and use an infrared transmitter to convert the data to an infrared signal. 红外检测器可以接收红外信号中携带的数据,而恰当的电路则可将数据安插到总线102上。 An infrared detector can receive the data carried in the infrared signal, and appropriate circuitry can be placed on the data bus 102. 总线102 将数据传送到主存储器106,处理器104从主存储器106中检索并寺丸行指令。 Bus 102 to transfer data to main memory 106, processor 104 retrieves from memory 106 and the main temple pill line instruction. 在由处理器104执行之前或之后,主存储器106接收的指令可以随意保存在存储设备110上。 Before or after execution by processor 104, instructions received by main memory 106 may be freely stored on storage device 110.

25计算机系统IOO还包括一个与总线102相连的通信接口118。 The computer system 25 further includes a communication interface IOO connected to a bus 102,118. 通信接口118提供了一个与网络链路120耦合的双向数据通信,其中网络链路120与本地网络122相连。 The communication interface 118 provides a two-way data communication coupled to a link with the network 120, network link 120 connected to a local network 122. 举例来i兌,通信4姿口118可以是一个综合业务数字网(ISDN )的网卡或是一个为相应类型的电话线路提供数据通信连接的调制解调器。 I versus way of example, the communication port 118 may pose 4 is an integrated services digital network (ISDN) card or a modem to provide a data communication connection to a corresponding type of telephone line. 作为另一个实例,通信接口118可以是一个局域网(LAN)网卡,它向兼容的LAN提供数据通信连4妄。 As another example, communication interface 118 may be a local area network (LAN) card, which provides a data communication connection to jump 4 compatible LAN. 此外还可以实施无线链路。 Further embodiments may also be a wireless link. 在4壬<可一种这类实施中, 通信接口118都会收发电、电磁或光信号,这些信号传送的是那些代表不同类型信息的数字数据流。 4-nonyl <may be one such embodiment, the communication interface 118 will receive power, electromagnetic or optical signals that are transmitted are those digital data streams representing various types of information.

网络链路120通常经由一个或多个网络而向其它数据设备提供数据通信。 Network link 120 typically provides data communication to the other data devices via one or more networks. 举例来i兌,网络链路120可以通过本地网络122来将一个连接提供给主机124或是互联网服务供应商(ISP ) 126运作的数据设备。 I against the way of example, network link 120 may be a connector 122 provided to the host 124 or Internet service provider (ISP) 126 through the operation of the local network device data. ISP126进而又经由现在通常称为"互联网,,128的全球分组数据通信网络来提供数据通信业务。本地网络122和互联网128 都使用了传送数字数据流的电、电》兹或光信号。经由不同网络的信号以及网络链路120上经由通信接口118的信号携带了往返于计算机系统100的数字数据,这些信号即为传送信息的载波的示范性形式。 ISP126 via turn now commonly referred to as "Internet ,, global packet data communication network 128 to provide data communication services Local network 122 and Internet 128 both use electrical transmitting digital data stream, an electrical" hereby or optical signals via different networks and the signals on network link 120 and carry the digital data to and from computer system 100 via communication interface 118, which signals that is exemplary forms of carrier waves transmitted information.

计算才几系统100可以经由一个或多个网络、网络链路120以及通信接口118发送消息并接收数据,其中包括了程序代码。 Only a few computing system 100 via one or more networks, network link 120 and communication interfaces 118 transmit messages and receive data, including program code. 在互联网实例中,服务器130可以经由互联网128、 ISP 126、本地网络122 以及通信接口118来发送一个用于应用程序的被请求码。 In the Internet example, server 130 may ISP 126, local network 122 and communication interface 118 to transmit a requested code for an application program through Internet 128,. 根据本发明,这种下载得到的应用是为这里所述的混合共享磁盘/无共享数据库系统提供的。 According to the present invention, the downloaded application such as described herein is mixed shared disk / shared nothing database system is provided.

接收到的代码可以按照接收时的原状由处理器104执行和/或存入存储设备IIO或其它非易失存储器而供稍后执行。 The received code may be executed by the processor 104 according to the undisturbed and / or stored in the storage device IIO or other non-volatile memory for later execution and the time of reception. 这样,计算机系统IOO可以得到栽波形式的应用码。 Thus, the computer system can be planted IOO application code in the form of waves.

26这里所述的在不同数据库服务器之间划分一个数据库所有权而对数据库访问加以控制的方法是在一个计算机系统上实施的,尽管根据本发明的一个方面,对某些"无共享"》兹盘数据的访问将会受到软件约束,但是对该计算机系统而言,可以从节点提供全部磁盘的共享磁盘访问,也就是说,它是一个可用于严格的共享》兹盘访问的系统。 Dividing a database ownership between different database servers and controlled database access method 26 described herein is implemented on a computer system, even though according to one aspect of the invention, certain "shared nothing" "hereby disc access to the data will be subject to software constraints, but in terms of the computer system can provide shared disk access to all disks from the node, that is, it is a shared system for strict hereby disk access. "

所有权组 Ownership Group

根据本发明的一个实施例,将一个数据库(或其某些部分)划分为所有权组。 According to one embodiment of the present invention, a database (or portions thereof) are divided into groups of ownership. 每个所有权组都被指派了一个或多个数据库服务器,以此作为所有权组的所有者。 Each ownership group is assigned one or more database servers, as the owner of the ownership group. 指派为所有权组所有者的数据库服务器被看作是从属于该所有权组的所有数据项的所有者。 Assigned to the database server ownership group is seen as belonging to the owner of all data items the ownership group of owners. 也就是说,这些数据库服务器准许直接访问该所有权组内部的数据项,而其他数据库服务器则不能直接访问这些数据项。 In other words, these database servers permit direct access to the interior of the ownership group of data items, while other database servers can not access these data items directly.

根据一个实施例,经常受到同时访问的数据项归合到相同的所有权组,以此确保它们由相同的数据库服务器所拥有。 According to one embodiment, the normalized data items frequently accessed simultaneously by the groups bonded to the same ownership, in order to ensure that they are owned by the same database server. 所有权组是 Ownership group is

通过将相关数据项组合^L为一个原子单元而对一组相关数据项进行操作的。 And operated by the relevant data item combinations ^ L as an atomic unit of a group of related data items. 举例来说,通过将一个所有权组的所有权从第一数据库服务器转移到第二lt据库服务器,该所有权组内部的所有数据项的所有权都可从第一数据库服务器转移到第二数据库服务器。 For example, a database server, the ownership of the ownership of the internal set of all data items can be transferred via a proprietary set of ownership transferred from the first to the second lt database server from the first server to a second database server database.

混合数据库系统 Hybrid database system

图2是一个对根据本发明一个实施例的混合数据库系统架构进行描述的冲匡图。 FIG 2 is a diagram for punch Peter described hybrid database system architecture according to one embodiment of the present invention. 图2包括三个节点202、 204和206,其上分别运行着数据库服务器208、 210和212。 Figure 2 contains three nodes 202, 204 and 206, respectively, running on the database server 208, 210 and 212. 数据库服务器208、 210和212 分别与緩冲存储器220、 222和224相关联。 Database servers 208, 210 and 212 respectively and the buffer memory 220, 222 and 224 are associated. 每个节点202、 204和206都与系统总线218相连,以使数据库服务器208、 210和212能够直4妄访问4呆存在这两个万兹盘214和216上的凄t据库250内部的数据。 Each node 202, 204 and 206 are connected to the system bus 218, so that the database server 208, 210, and 212 can direct access jump 4 4 presence of these two stay Wan hereby tray 214 and data 216 t on desolate internal database 250 data.

包含在磁盘214和216上的数据从逻辑上划分为所有权组230、 232、 234和236。 Data contained on the disk 214 and 216 is logically divided ownership group 230, 232, 234 and 236. 才艮据本发明的一个实施例,每个所有权组包含一个或多个表空间(tablespace )。 According to a Gen only embodiment of the invention, each group comprising one or more of the ownership table space (tablespace). 表空间是一个或多个数据文件(datafile)的集合。 Table space is the set of one or more data files (datafile) a. 然而,本发明并不局限于任何一种特定的划分颗氺立度,并且可以与更大或更小范围的所有4又组一起^f吏用。 However, the present invention is not limited to any particular division of Li Shui particles, and may be with a larger or smaller ranges and all 4 groups ^ f with officials.

才艮据一个实施例,每个所有权组都^皮指定为共享》兹盘所有权组或是无共享所有4又组。 According to one embodiment, only Burgundy, each ownership group are designated as shared ^ skin "Have disk or shared-ownership group and all 4 groups. 指定为无共享所有权组的各个所有权组都指派了一个可用数据库服务器,以此作为它的所有者。 Designated as shared-ownership group ownership of each group are assigned a usable database server, as its owner. 在图2描述的系统中,所有权组230是服务器210拥有的无共享所有权组,所有权组232是一个共享磁盘所有权组,所有权组234是服务器212拥有的无共享所有^又组,所有权组236是服务器208拥有的无共享所有权组。 In the system described in FIG. 2, the title group 230 is a server not shared ownership group owned by 210, title group 232 is a shared disk ownership group, title group 234 is the server 212 has no share all ^ and group ownership of the group 236 is server 208 has no share ownership group.

由于所有权组230是服务器210拥有的无共享所有权组,因此只允i午月良务器210直接^访问所有冲又组230内部的凝:据(Dl )。 Since the ownership group 230 is a server shared-ownership group that owns the 210, so i only allowed the afternoon on the good works ^ 210 direct access to all red and condensate inside the group 230: According to (Dl). 任何试图访问所有权组230数据的其它服务器通常需要向服务器210发送请求,要求服务器210作为请求服务器的代表来执行预期的数据访问。 Any attempt to access other servers title group 230 transmits data to the server 210 typically requires request to the server 210 to perform the desired data access requests to the server as a representative. 同样,所有权组234和236也是无共享所有权组,并且只能被它们各自的所有者直接访问。 Similarly, the ownership group 234 and 236 is no shared ownership group, and can only be accessed directly to their respective owners.

由于所有权组232是一个共享磁盘的所有权组,因此任何数据库服务器都可直接访问其中包含的数据集。 Since the title group 232 is a set of disks shared ownership, any database server can directly access the data sets contained therein. 如图2所示,每个数据库服务器都可在其緩冲存储器中包含该数据(D2)的一个副本。 2, each database server can contain a copy of the data (D2) in its buffer memory. 并且使用了一个分布式锁定管理器来对共享数据访问进行协调。 And it uses a distributed lock manager to coordinate access to shared data.

根据一个实施例,该数据库系统包括一种机制,用于将某个特定所有4又组从共享^兹盘动态改变为无共享,反之亦然。 According to one embodiment, the database system includes a mechanism 4 for all but a specific group from the shared disk ^ hereby no shared dynamically change, and vice versa. 举例来i兌, I against the way of example,

28如果某个特定无共享数据集受到频繁的读访问(读热点),那么可以将其所属的所有权组从无共享转换为共享磁盘,从而将该数据转换为共享磁盘。 28 If a particular set of shared-data access by frequent reading (read hot), so it can be shared ownership group belongs never converted to shared disk, which converts the data to a shared disk. 同样,如果某个共享磁盘的数据集受到频繁的写访问(写热点),那么可以将包含该数据的所有4又组改成一个无共享所有权组,并将该所有权组的所有权指派给一个数据库服务器,由此将该数据转换成无共享数据。 Similarly, if a shared disk data sets by frequent write access (write hot spots), you can turn all four groups will contain no data into a shared ownership group and assign ownership of the ownership group to a database server, whereby the data is converted into non-shared data.

才艮据本发明的一个方面,该数据库系统还包括一种机制,用于把无共享所有权组的所有权从一个节点重新指派给另一个节点。 According to a Gen only aspect of the invention, the database system further comprises a mechanism for the non-shared ownership rights from a group can be assigned again to another node node. 操作员可能需要该操作来改善负栽均衡,并且该操作也可自动进行, The operator may need to improve the operation of the plant negative balance, and the operation can be carried out automatically,

以便在节点Nl出现故障之后继续为访问Nl所拥有的无共享所有权组的数据提供支持。 In order to continue to provide support for shared-ownership data access Nl group owned by the node Nl after failure.

所有权 ownership

如上所述,提供了一种数据库系统,其中某些所有权组被指定为无共享所有权组,而某些所有权组则被指定为共享磁盘的所有权组。 As noted above, it provides a database system, some ownership groups are designated as shared-ownership group, and some ownership groups were designated as shared disk ownership group. 每个无共享所有权组都被指派了一个所有者。 Each shared-ownership group is assigned an owner. 对所有数据库服务器来说,无共享所有权组的所有权是已知的,因此,在需要对所有权组内部数据执行任务的时候,这些数据库服务器可以向所有权组的所有者发送请求。 For all database servers, without shared ownership ownership group is known, therefore, when you need to perform tasks on the internal data ownership group, the database server can send a request to the owner of the ownership group.

根据本发明的一个实施例,不同所有权组的所有权信息都保存在一个控制文件中,并且所有能够访问该数据库的^U居库服务器都准许访问该控制文件。 According to an embodiment of the present invention, ownership information with the ownership group are stored in a control file, and all access to the library server UN ^ U are granted access to the database control file. 每个数据库服务器都可将控制文件的一个副本存入緩存。 Each database server can control a copy of the file is stored in the cache. 在緩存中具有控制文件副本的情况下,数据库服务器可以确定该所有4又组的所有权,而不必始终承担与从》兹盘中读取所有权信息相关联的开销。 In the case of a cache copy of the control file, the database server can determine that all ownership and group 4, without having to always shoulder the cost associated with ownership information is read from the "Have disk.

图3是描述数据库服务器所执行的步骤的流程图,该数据库服务器希望得到既使用共享磁盘所有权組又使用无共享所有权组的 3 is a flowchart describing the steps performed by the database server, the database server wants to use both shared disk ownership group and use of a shared ownership group

29系统中的数据。 29 Data System. 在步骤300,数据库服务器确定预期数据所属的所有权组。 In step 300, the database server determines ownership group expected data belongs. 在步骤302,数据库服务器确定包含预期数据的所有权组的所有者。 In step 302, the database server determines the owner of the ownership group that contains the expected data. 如上所述,可以通过访问一个控制文件来执行步骤302, 其中可以将控制文件的一个副本存入与数据库服务器相关联的緩存。 As described above, a control file may be performed by accessing step 302, where a copy of the cache can be stored in the control file associated with the database server. 如果该所有斥又组是一个共享》兹盘所有4又组,那么所有数据库服务器都将被看作是该所有权组的所有者。 If all this is a shared group has denounced "Have disk and all 4 groups, all database servers will be regarded as the owner of the ownership group. 如果该所有权组是一个无共享所有权组,那么将会在控制文件中把某个数据库服务器指定为该所有^L组的所有者。 If the ownership group is a shared-ownership group, it will put a database server designated as the owner of all ^ L groups in the control file.

在步骤304,数据库服务器判定自身是否即为保持预期数据的所有4又组的所有者。 In step 304, the database server determines whether it is the owners expect to keep all the data and 4 groups. 如果(1)该所有4又组是一个共享石兹盘的所有权组,或者(2)该所有权组是一个无共享所有权组,并且控制文件中把该数据库服务器指定为无共享所有权组的所有者,那么该数据库服务器即为该所有权组的所有者。 If (1) that all four and ownership group is a shared set of stone hereby disk, or (2) the ownership group is a shared-ownership group, and a control file to the database server as the owner of the shared-ownership group , then the database server is the owner of the ownership group. 如果数据库服务器是保持预期数据的所有权组的所有者,那么控制将会传递到步骤310,在那里数据库服务器将会直接检索预期数据。 If the database server is expected to remain the owner of the ownership group of data, control is passed to step 310, where the database server will directly retrieve the expected data.

如果数据库服务器不是保持数椐的所有权组的所有者,那么控制将会传递到步骤306。 If the database server is not noted in the number of owners to maintain ownership group, then control is passed to step 306. 在步骤306,数据库"l务器向该所有权组的所有者发送一个请求,以使该所有者代表请求者来访问预期数据。在步骤308,数据库服务器从该所有权组的所有者接收预期数据。 At step 306, the database "l sends a service request to the owner group ownership, so that the owner is expected to access the data on behalf of the requestor. In step 308, the database server receives the expected data from the owner of the ownership group.

所有者集合 The owner of the collection

根据一个替换实施例, 一个所有权组并不局限于(1)只由仅仅一个数据库服务器所拥有(无共享),或者(2)由所有的数据库服务器所拥有(共享^兹盘)。 According to an alternative embodiment, the ownership of a group is not limited to (1) have only (non-shared) by the only one database server, or (2) have (shared disk hereby ^) of all database servers. 相反,所有权组还可以由可用数据库服务器中的任何指定子集所拥有。 In contrast, the group may also have ownership of any given subset of the available database server. 在这里,拥有某个特定所有权组的数椐库服务器集合称为所有权组的所有者集合。 Here, have a particular ownership group noted in the number of database server set is called the owner of the ownership group of the collection. 因此,无共享所有权组与所有者集合中只包含一个数据库服务器的所有权组等价,而共享磁盘所有权组则与所有者集合中包含了所有可用数据库服务器的所有权组等价。 Therefore, no shared ownership and group ownership group equivalence owner collection contains only one database server and shared disk ownership group contains a set of equivalent proprietary database of all available servers and owner collection.

当使用所有者集合来执行一个关于所有^又组的数据的任务时, 不属于该所有权组的所有者集合的数据库服务器将会向那些归属于该所有权组的所有者集合的数据库服务器中的一个服务器发送请求„响应于该请求,请求接收者直接访问该所有权组的数据并且执行所请求的任务。由所有权组内部写热点产生的争用只会在那些归属于所有权组的所有者集合的数据库服务器之间产生。 The database server when the database server to perform a task on a group of all ^ and use of data collection when the owner does not belong to the owner of the collection will be in the ownership group of the owner of the ownership group of those belonging to the collection of a server sends a request "in response to the request, requesting recipients to directly access data in the ownership group and perform the requested task. contention hot spots generated by the write internal database ownership group with only those belonging to the owner of the collection of the ownership group generated between servers.

改变所有4又组的所有权 All changes in ownership and group 4

如上所述,把所有权组从无共享变成共享磁盘或从共享磁盘变成无共享,这是合乎需要的。 As mentioned above, the group never shared ownership becomes a shared disk or shared-nothing becomes a shared disk, which is desirable. 这种变化可以响应于读写热点检测而 This change in response to a write focus detection

净皮自动启动,也可以手动启动(例如响应于勒:才居库管理员发出的命令)。 The net skin starts automatically, you can manually start (for example, in response to Le: Habitat command was issued by the database administrator).

可以使用不同技术来把所有权组从一个所有者集合("源所有者集合")转移到另一个所有者集合("目的所有者集合")。 Different techniques may be used to set the ownership of the group ( "Source Owner set") to another set of the owner from one owner ( "owner object set"). 图4是一个描述才艮据本发明实施例来改变所有4又组的所有者集合而执行的步骤的流程图。 FIG 4 only is a Gen-described embodiments of the invention, according to all the steps of the flowchart of changing a set of the owner of the group 4 and executed.

参考图4,在步骤400, 一个"禁止改变"的消息广播到所有可用的数据库服务器。 Message Referring to FIG 4, in step 400, a "No change" is broadcast to all available database server. 该禁止改变消息命令数据库服务器停止对那些所有者集合将会变化的所有权组("转移中的所有权組")内部的数据做出前向改变。 The message command prohibits changing the database server stops ownership group ( "group transfer of ownership in") collection will change for those inside to make the owner of the data before the change. 前向改变是那些创建一个先前并不存在(也就是创建一个新的"当前"版本的数据项)的版本的改变。 Forward to create a change for those who previously did not exist (that is, create a new "current" version of the data item) version change. 另一方面, 后向改变则是那些导致重新创建先前已有版本的数据项改变。 On the other hand, after the change are those that re-creates the data item to change the pre-existing version.

31在步骤402,负责改变所有权组的所有者集合("所有者改变机制")的那部分数据库一直等待,直到所有改变了转移中的所有权组的事务都已提交(commit)或回滚(roll back )。 31 In step 402, the owner is responsible for changing the ownership group of the set ( "owner change mechanism") that part of the database waits until all transactions changed ownership group of metastasis have been submitted to (commit) or rolled back (roll back).

由于不再允许对所有权组进4亍前向改变,因此,那些在步骤400 Since no longer allowed to change the ownership into 4 groups before the right foot, so that at step 400

事务将会回滚。 Transaction is rolled back. 由于步骤400仅仅阻止那些对转移中的所有权组的前向改变,因此不会防止数据库服务器回滚那些它们已对转移中的所有权组^L出的改变。 Because only 400 steps to stop those changes before the transfer of ownership of the group, it will not prevent the database server to roll back those changes they have to transfer the ownership group ^ L out.

不幸的是,可能会需要大量开销来判定哪些事务更新了转移中的所有权组。 Unfortunately, it may need a lot of overhead to determine which transactions update the transfer of the ownership group. 因此提供了本发明的一个实施例,其中数据库系统并不尝试追^宗那些更新了转移中的所有冲又《且以内的凄t梧的事务。 Accordingly the present invention provides an embodiment wherein the database system does not attempt to recover all those updates ^ were washed and "sad t Wu and within the transfer transaction. 然而,在不对该信息进行追踪的情况下,必须4叚设任何一个准许访问转移中的所有权组数据并始于步骤400之前的事务都改变了转移中的所有权组内部的数据。 However, without this information for tracking, 4 Jia must set up a data set of any ownership transfer in and grant access before the transaction began in step 400 has changed the internal data transfer of the ownership group.

基于这个假设,步骤402要求改变机制的所有者一直等待,直到(1)可能已经访问了转移中的所有权组数据,以及(2)始于步骤400之前的所有事务提交或回滚。 Based on this assumption, step 402 requires the owner to change the mechanism waits until (1) may have access to the data transfer in the ownership group, and commit all transaction or rollback before 400 (2) begins with step. 通常,只有那些在归属于变换所有权组的源所有者集合的数据库服务器中运行的事务才有可能访问转移中的所有4又组的凄t据。 Typically, only those transactions running on the database server attributable to transform the source ownership group of the owner of the collection will it be possible to access all sad t 4, according to another group of metastasis. 因此,如果转移中的所有权组是共享磁盘的,那么改变机制的所有者必须一直等待,直到所有数据库服务器中始于步骤400之前的所有事务都已提交或是回滚。 Therefore, if the transfer of the ownership group is a shared disk, then the owner must change the mechanism waits until all transactions until all of the database server 400 begins at step have been committed or rolled back. 如果转移中的所有;K组是无共享的,那么改变才几制的所有者必须一直等待,直到拥有转移中的所有权组的数据库服务器的所有事务都已提交或者回滚。 If all metastasis; K group is not shared, then the change must be made before a few owner waits until all transactions have transferred ownership of the database server group have been committed or rolled back. 需要注意,其中包含了那些在其他节点发起并已创建了转移中的所有权组的本地子事务的用户事务。 It should be noted, which contains the user who initiated the transaction and the other node has created a local ownership group of sub-transactions in the transfer.

在有可能更新了转移中的所有权组内部数据的全部事务都已提交或回滚的时候,控制将会前进到步骤404。 There may be updated in the internal affairs of all data transferred in the ownership group have been committed or rolled back when control will proceed to step 404. 在步骤恥4,改变机制的所有者通过更新最小单元操作中的控制文件来改变转移中的所有权组的所有者集合。 In step 4 shame, the owner of the mechanism by changing the file update control minimum unit operation to change the owner of a set of transfer of ownership of the group. 举例来说,该指示变化可以使转移中的所有权组从无共享所有权组转移到共享^兹盘的所有权组,反之亦然。 For example, this can indicate a change in the transfer of ownership from the non-group share ownership transfer the group to the shared disk ^ hereby group ownership, and vice versa. 作为选择,指示变化可以只改变拥有一个无共享所有权组的数据库服务器,而不改变该所有权组的类型。 Alternatively, the change may indicate no change has only a shared database server ownership group, without changing the type of the ownership group.

在改变控制文件,使之反映转移中的所有权组的新的所有者集 Change control documents to reflect the new ownership group of the owner of the collection of metastasis

合时,控制将会前进到步骤406。 Timely, control will proceed to step 406. 在步骤406, 一个"刷新緩存" 的消息将会发送到所有的可用数据库服务器。 In step 406, a "Refresh cache" message will be sent to all the available database server. 一旦接收到刷新緩存的消息,每个数据库服务器将会无效掉緩存中包含的控制文件副本。 Upon receipt of the message refresh the cache, each database server will invalidate cached copy of the control file contains. 因此,当数据库服务器随后需要检查控制文件来确定所有权组的所有权时,它们会从持久存储器中检索更新版本的控制文件。 Therefore, when the database server then you need to check the control file to determine the ownership of the ownership group, they will retrieve the updated version of the control file from persistent storage. 由此这些数据库月l务器将会了解到转移中的所有权组的新的所有者集合。 Whereby these databases month l Service is learned that the new owner will be a collection of transfer of the ownership group.

对所有权变化的调整 For adjustments and changes of ownership

在频繁使用某个查询的时候,该查询通常会保存在数据库内部。 When frequent use of a query, the query is usually stored in the internal database. 在将所保存的查询初始提交到数据库系统的时候,大部分数据库系统都会为所保存的查询产生一个执行计划,而不是在每次使用所保存查询的时候重新计算一个运行方案。 When the saved initial query submitted to the database system, most database systems will produce an implementation plan for the saved query, rather than recalculated every time a program run using the saved query. 一个查询的执行计划必须顾及包含该查询所访问数据的所有权组的所有权。 A query execution plan must take into account the ownership of the ownership group that contains the query to access data. 举例来说,如果该查询指定了对于某个特定数据库服务器专门拥有的所有权组中的一个数据项的更新,那么该查询的执行计划必须包括将该更新操作发送到该特定数据库服务器。 For example, if the query specifies a data item update for a particular database server ownership group has specialized in, then the query execution plan must include the specific update operation is sent to the database server.

然而如上所述,为了改变所有权组的所有权而提供了一种机制。 However, as mentioned above, in order to change the ownership of the ownership group and provides a mechanism. 这种所有权变化可以出现在已经为某个特定所保存查询生成了执行计划之后。 This change of ownership may occur after you have saved for a particular query execution plan is generated. 因此,这些执行计划可能需要某些数据库服务器对它们不再拥有的所有权组内部数据执行操作。 Therefore, the implementation of these plans may need to perform some operation data ownership group internal database server they no longer have. 根据本发明的一个实施例,那些要求对不为其所有的所有权组内部数据执行操作的数据 According to one embodiment of the present invention, that requirement is not for internal data of all data sets to perform operations ownership

33库服务器向请求该操作的进程返回一个"所有权错误,,的消息。响应于接收到一个所有权错误消息,将会为该导致错误的查询产生一个新的执行计划。新的执行计划考虑到了当前版本的控制文件所指示的所有权组的当前所有权。 33 database server returns an "error ,, ownership of the operation to process the request message. In response to receiving a title error message will generate a new execution plan for the query that caused the error. The new implementation plan takes into account the current the current ownership group ownership of version control file indicates.

控制文件的管理 Management control file

如上所述使用了一个原子操作来更新控制文件,以便改变所有 As described above we use an atomic operation to update the control file, in order to change all

权组的标志(步骤404 )。 Flag (step 404) of the right group. 在这里可以使用不同机制来确保该操作是最小单元的操作。 Herein may use different mechanisms to ensure that the operation is an operation of a minimum unit. 举例来iC,如图5所示,才艮据本发明的一个实施例,控制文件包4舌一个位图和一系列都:据块对。 IC way of example, as shown in Figure 5, only Gen embodiment of the present invention according to one embodiment, the control package of the tongue 4 and a series of bitmaps are: data block pair. 4立图512中的每个比特都与一个lt据块对相对应。 FIG 4 stand 512 with each bit of a data block corresponding to lt.

在4壬4可给定时间, 一个潮:据块对中只有一个凄史据块包含当前数据。 4 may be 4-nonyl given time, a tide: only one data block to data block contains the current history desolate data. 与数据块对相关联的比特值指示了对应数据块对的两个数据块中哪一个保持当前数据。 Bit value of the data block associated with the data indicative of the two blocks corresponding to data blocks for which one of the current data retention. 举例来说,比特502与包含数据块506和508的数据块对504相关联。 For example, 502-bit data of the data block 506 and block 508 to 504 comprising associated. 比特502的值(例如"0")代表数据块506是数据块对504内部的当前数据块。 502 bit value (e.g. "0") represents a data block is a data block 506 of the current data block 504 inside. 比特502的值可以变为"1",这表示数据块508的数据是当前数据(由此数据块506中的数据不再有效)。 Value of the bit 502 may be changed to "1", which indicates the data block 508 is the current data (block 506 whereby the data is no longer valid).

由于一个数据块对的非当前数据块的数据被认为是无效的,因此可以将数据写入非当前数据块,而不改变控制文件的有效内容。 Since the data of a data block of the non-current data block is considered invalid, it is possible to write data to the non-current data block, without changing the effective content of the control file. 实际上只有在位图512的比特值发生改变的时候,控制文件的内容才会改变。 In fact, only the bit value of the bit 512 of FIG changed when the contents of the control file will change. 因此,作为原子变化的预备步骤,数据块对504中当前数据块506的内容可以加载到存储器中,也可以被修改,还可以存入数据块对504的非当前数据块508。 Thus, as a preliminary step change atoms, the content of the data block 504 in the current data block 506 may be loaded into the memory, may also be modified, the data may be stored in block 508 to block 504 of the non-current. 在执行了这些预备步骤之后, 该变化可以通过改变与数据块对504相对应的位图512内部的比特502的值而被自动进行。 After performing these preparatory steps, the change may be by changing the data blocks 504 corresponding to the value of the internal bitmap 512 bit 502 is automatically performed.

34这仅仅是自动执行改变的技术的一个实例。 34 This is only one example technique performed automatically changed. 其它技术也是可以接受的。 Other techniques are also acceptable. 因此,本发明并不局限于任何一种用于自动执行改变的特有技术。 Accordingly, the present invention is not limited to any specific technique for automatically performing changes.

在所有相j且之间移动数才居项 Only living relative movement between the number of items and all j

一种改变诸如表空间这种数据项所有权的方法是改变数据项归属的所有权组的所有者。 A method of modifying the ownership of such items, such as table space is to change the ownership of the owner of the data item belongs group. 另一种改变数据项所有权的方法是把数据项重新指派给不同的所有权组。 Another way to change the ownership of the data item to a data item is reassigned to a different group ownership. 举例来说,通过从指派给服务器 For example, assigned by the server

A的所有权组中移除表空间A并将其安插在指派给服务器B的所有权组中,可以将表空间A的所肴者从服务器A改为服务器B。 A set of ownership removed and placed tablespace A set of ownership which is assigned to the server B, the dishes by the table space A can be changed to the server from the server B. A

才艮据本发明的一个实施例,所有权组的全体成员都保存在数据库内部的一个数据字典中。 According to the present invention, a Gen only embodiment, the ownership of all members of a group are stored in the internal database data dictionary. 因此,要从第一所有权组将一个数据项移动到第二所有权组,那么必须更新数据字典内部的第一和第二所有权组的全体成员的信息。 Therefore, from the first ownership of a set of data items will be moved to the second ownership group, you must update the information of all members of the ownership group of the first and second internal data dictionary. 在改变数据项所有权组时所涉及的不同步骤与那些用于改变所有权组的所有者集合的上述步骤相似。 Different steps in changing the ownership of the group of data items are similar to those according to the above-described collection step for changing the ownership of the owner of the group. 具体地说,对于正在转移的表空间("转移中的表空间,,)的访问是被禁止的。然后,所有权变化机制将会等待锁定在数据项上的所有事务(或其组件)回滚或者提交。 Specifically, for table space is being transferred ( "the transfer of table space,,) access is prohibited. Then, change the ownership of the lock mechanism will wait for all transactions (or its components) in the data item rollback or submit.

一旦锁定在数据项上的所有事务都已提交或者回滚,那么数据字典将^皮修改,以便指示该数椐项的新的所有权组。 Once locked on the data items that all transactions are committed or rolled back, then the data dictionary will modify ^ skin to indicate the new ownership group noted in the number of entries. 然后控制文件将被修改,以便指示数据项移至的所有权组的所有者集合即为数据项的所有者集合。 Then the control file will be modified to indicate the owner of the data item to the ownership group is the owner of a collection of data items in the collection. 该改变自动允许目标所有者访问该数据项。 This change automatically allow the owner to access the target data item. 如果所有^J且正在所有4又变化之中,那么控制文件将^皮更新,以^^指示该数据项处于一个"移动延迟"的状态中。 If among all J ^ 4 and all variations and is, then the updated control file transdermal ^, ^^ to indicate that the data item is a "delay movement" state.

对于数据项所属的所有权组的变化有可能会改变数据项所有者,也有可能不改变数据项所有者。 For a change of ownership belongs to the group of data items may change the owner of the data item, there may not alter the data item owner. 如果源所有权组的所有者集合与转移中的所有权组的所有者集合相同,那么,在数据项从源所有权组移动到转移中的所有权组的时候,数据项的所有者不会改变。 If the owner of the ownership group of the source collection and transfer of the owner of the ownership group of the same set, then the data items from the source ownership group to move when the transfer of ownership of the group, the owner of the data items will not change. 另一方面,如果源所有权组的所有者集合不同于转移中的所有权组的所有者集合,那么,在把数据项从源所有权组移动到转移中的所有权组的时候,数据项的所有者将会改变。 On the other hand, if the ownership of the owner of the source is different from the owner of the group set of the set of transfer of ownership of the group, then the time of transfer of ownership of the group, the owner of the data item is moved from the source data item to the group ownership It will change.

特定所有权的变化条件 Changes in ownership of certain conditions

根据一个实施例,提供了某些技术方法来处理以下情况,其中: According to one embodiment, there is provided a method to deal with some of the following techniques, wherein:

(1 )在归属于一个所有权组的数据项正在转移到另外一个所有权 (1) in the ownership of a data item belonging to the group is transferred to another ownership

组的时候,尝试改变该所有权组的所有者集合;以及(2)当目的 When the group attempts to change the ownership of the owner of the group set; and (2) when the aim

地所有权组正在改变其所有者集合的时候,尝试将数据项转移到另外一个所有^又组。 When the ownership group is changing its owner collections, try to transfer all the data items to another ^ and groups.

为了^企测这些状态,本发明的一个实施例在控制文件内部为归属于一个所有权組的每个数据项(例如表空间)提供了一个或多个状态标志。 To measure these enterprises ^ state, a present embodiment of the invention provide one or more status flags is assigned to each data item (e.g. tablespace) ownership of a group within the control file. 举例来iJt,可以^使用一个标记来指示^:据项归属的所有权组是否处于指派新的所有者的过程中。 Way of example iJt, can use a flag to indicate ^ ^: whether the data item in the group home ownership assignment process in the new owner. 同样, 一个标记可以指示数据项正处于转移到不同所有权组的过程中。 Similarly, a flag may indicate that the data item is in the process of transferring ownership to a different group.

在尝试改变所有权组的所有者集合的时候,所有权变化机制将会检查归属于所有权组的数据项的状态标志,以便确定归属于该所有权组的任何数据项是否处于变换到不同所有权组的过程中。 In the attempt to change the ownership group of the owner of the collection when the ownership change mechanism will check the status flag data items belonging to the ownership group to determine any data item attributable to the ownership group is in the process of converting to a different ownership group in . 如杲归属于该所有权组的任何一个数据项处于转移到另外一个所有权组的过程中,那么将会停止尝试改变该所有4又组的所有者集合。 Gao as a data item belonging to any of the ownership group is in the process of transferring ownership to another group, it will stop all attempts to change the owner and group 4 of the set. 如果没有一个归属于该所有权组的数据项处于转换到不同所有权组的过程中,那么归属于该所有权组的数据项的状态标志指示数据项归属的所有4又组的所有权正处在转移之中。 If you do not belong in a data item in the ownership group of the transition to a different ownership group, then the status flag data items belonging to the ownership group 4 and group ownership of all data items indicating the transfer of ownership is in among . 并且还将一个消息发送到不同的数据库服务器,以便无效掉它们緩存的控制文件版本。 And a message will be sent to a different database server, to invalidate the control files are cached version. 由此确保了这些数据库服务器了解到新的状态标志值。 This ensures that the database server understands the new state flag.

36当尝i式把凄tl居项移动到不同所有^又组的时〗,美,数|居项的状态标志将被检查,以便确定目的所有权组是否处于改变其所有者集合的过程中。 36 When the mournful style taste i tl ranking items moved to a different all the time ^〗 another group, the United States, the number | Habitat term state flag will be examined in order to determine the purpose of the ownership group is in the process of changing its owner collection. 根据一个实施例,这个检查的执行发生在修改数据字典使之反映新的数据项所有权组之后,以及更新控制文件而使新的所有权组的所有者能够访问该数据项之前。 According to one embodiment, this occurs after a check is performed to modify the data dictionary to reflect the new data item group ownership, and file update control ownership of the new owner of the group can access the data item before. 如果数据项归属的所有权组处在改变其所有者集合的过程中,那么用于控制文件中的数据项的状态标志将被设置成指示一个"移动延迟"的状态。 If the data item in the group home ownership change procedure set its owner, then the status flag data item for controlling the file will be set to indicate a "delayed movement" state. 另外,整个数据库中的"移动延迟"标记将被设定为指示该数据库中包含了某些处在移动延迟状态中的数据项。 Further, the entire database "Mobile delay" flag is set to indicate that the database contains certain data items in a mobility retardation state.

在完成了转移所有权组成员的操作的时候,执行变换的进程将会更新状态标志,以便指示该所有权组不再处于所有权转移的过程中。 Upon completion of the transfer of ownership operation group members when performing transformation process will update the status flag to indicate that the group is no longer in the process of ownership transfer of ownership. 另外,该进考呈还清除了在所有4又组的所有4又转移过程中移动到该所有权组的〗壬何数据项的"移动延迟"标志。 Further, in addition to the intake test was moved to pay off "Mobile delay"〗 nonyl any data item in the group ownership of all the transfer process and all 4 groups and 4 flag.

故障排除 Troubleshooting

在所有权改变的过程中有可能会发生故障。 It is possible to change the ownership of the process fail. 该故障有可能是 The fault may be the

"进程死亡(progress death ),,或是"服务器宕机(server death ),, 的结果。 "Process Death (progress death) ,, or" results server goes down (server death) ,, of. 在涉及所有权改变的某个进程出现故障的时候,将会出现一个进程死亡。 When a process involving change of ownership of a failure, there will be a process of death. 在整个数据库服务器发生故障的时候,将会出现一个服务器宕机。 When the entire database server failure, a server crash will occur. 在存在这两种故障类型的情况下,持久存储器上仍未保存的所有改变也许会丟失。 In the presence of these two types of failures, all lasting change has not saved on the memory may be lost. 在这种故障之后,有必要将数据库返回到一种一致状态。 After such a failure, it is necessary to return the database to a consistent state.

根据本发明的一个实施例,从进程死亡中恢复是通过使用状态对象(state object)来执行的。 According to one embodiment of the present invention, recovery from the process of death is performed by using state objects (state object). 状态对象是一种数据结构,它被分配在了与进程所属的数据库服务器相关联的存储器区域中。 State object is a data structure that is allocated in the memory area of ​​the database server process belongs associated. 在执行一个动作之前,进程会更新状态对象,以便指示它要执行的动作。 Before executing an action, process updates the state of the object, so that it indicates the action to be performed.

如果进程死亡,那么服务器内部的另一个进程(例如"进程监视器")将会调用该状态对象的一种方法(一个"清除程序")而把数据库返回到一个一致状态。 If the process dies, another process within the server (for example, "Process Monitor") will invoke a method of the state of the object (a "clear program") and the database back to a consistent state.

在进程出现故障之后执行来进行清除的特殊动作取决于失效进程所执行的操作以及失效进程死亡在其失效之前已经执行到的程度。 Special action after the implementation process fails to remove depends on the operating process failure and failure process executed before the death of the extent of the failure has been executed to. 根据一个实施例,在所有权组的所有权改变过程中,进程故 According to one embodiment, the process of the change of ownership of the ownership of the group, so the process

障是如下处理的: Barrier is a process of:

如果执4亍所有权变化的进程在其改变最终控制文件之前失效, 那么初始的所有者将会作为所有权组的所有者而得到恢复。 If executed right foot 4 ownership change process before it changes, the final control file fails, the original owner will be as the owner of the ownership group and restored.

如果执行所有权变化的进程是在改变最终控制文件之后但是在删除状态对象之前失效的,那么新的所有者仍然是该所有者,并且状态对象将被删除。 If you perform the change of ownership of the process is in control after changing the final document but before you delete the state object is invalidated, then the new owner continues to be the owner, and the state of the object will be deleted.

在将数据项从一个所有权組转移到另一个所有权组的时候,进禾呈故障是如下处理的: When the ownership of items from one group to another group ownership, into the fault was Wo is a process of:

如果执行转移的进程是在改变数据字典之前失效的,那么数据项的初始所有者将会作为数据项所有者而得到恢复。 If the branch before failure is in the process of changing the data dictionary, then the initial owner of the data items as data items will be the owner and restored.

如果执行转移的进程是在提交了涉及字典的变化之后但是在最终控制文件发生改变之前失效的,那么进程监视器将会结束移动并对控制文件进行适当改变。 If the branch is in the process but before the final failure of the control file is changed after the submission of the changes involves the dictionary, then the process will end mobile monitor and control files appropriately changed. 如果该所有权组处在所有权变化中, 那么数据项处于"移动延迟"状态中。 If the ownership of the ownership changes in the set, then the data item is in the "Mobile delay" state.

如果执行转移的进程是在最终控制文件改变之后但是在删除状态对象之前失效的,那么进程监视器将会删除状态对象。 If the branch is in the process but before you delete the state object is invalidated after the final control file change, then the process monitor will delete the state object.

38服务器宕机 38 server downtime

虽然数据库服务器宕机,但是并没有向宕机服务器单独拥有的所有权组的数据提供访问。 While the database server is down, but does not provide access to data ownership group has a separate server downtime. 因此,根据本发明的一个实施例,服务器宕机是一个事件,它触发了一个自动的所有权变化,其中,出现故障的服务器单独拥有的全部所有权组都指派给了新的所有者。 Thus, according to one embodiment of the present invention, the server downtime is an event, which triggers an automatic change of ownership, wherein individual servers have failed full ownership groups were assigned to a new owner.

在服务器故障之后执行的用于进行清除的特殊动作取决于数据库服务器正在执行的操作以及服务器宕机前执行了多少所有权转移。 Special actions for execution after a server failure is cleared before the operation depends on the database server is down and the server is performing much of the implementation of the transfer of ownership. 4艮据一个实施例,所有才又组的一个所有冲又变4匕过程中的服务 It is all a punch 4 Gen one embodiment, all the groups has changed before and during service dagger 4

器故障是如下处理的: Failure is a process of:

如果在改变最终控制文件之前源数据库服务器已经宕机,那么所有权组将会指派到另一个线程,控制文件中的状态信息将会更新,以1更指示所有4又组不再处于转移之中。 If the source database server has been down before changing the final control file, then the ownership group will be assigned to another thread status information in the control file will be updated to indicate that all 4 and 1 more groups no longer in the transfer.

如果目标数据库服务器宕机,那么(l)执行转移的进程将会检测到实例(instance)失效并且终止转移,或者(2)在宕^几服务器的恢复过程中,所有权组将会从宕机服务器重新指派到另一个服务器。 If the target database server is down, then the process (l) transfer of the implementation will be detected instances (instance) failure and stop transfer, or (2) in the recovery process dang ^ several servers, the ownership group will be down from the server reassigned to another server.

在将数据项从一个所有^J且转移到另一个所有4又组时出现的服务器故障是如下处理的: In all of the ^ J from a data item and transferred to another server failure occurs when all but 4 treatment groups are as follows:

如果源服务器是在字典变化之前宕才几的,那么在恢复过程中, 新的所有者将被指派给源所有权组并会清除数据项的转移标记。 If the source server is in the dictionary before dang only a few changes, then during the recovery process, the new owner will be assigned to the source ownership group and clears the marks transferred data items.

如果源服务器是在改变字典之后但是在改变最终控制文件之前宏机的,那么在源服务器的恢复过程中会为数据项指派恰当的所有者或是将其标记为移动延迟,由此结束移动操作。 If, after the source server is changed to change the dictionary but before final control machine macro file, then the recovery process will be assigned to the source server in the proper owner of the data item or to mark it as a delay moved, thereby moving the end of the operation . 如果目标服务器宕机,并且已经改变了最终控制文件,那么数据项将会标记为"移动延迟"。 If the target server is down, and have changed the final control file, then the data item will be labeled "move delayed." 在宕机服务器的恢复过程中,转移中的所有权组的所有权将被重新指派,移动延迟标记将被清除。 During recovery server downtime, ownership transfer of the ownership of the group to be reassigned, the mobile delay flag is cleared.

减少所有权变化过程中的停机时间 Reduce downtime in the process of ownership change

如上所述,图4描述的步骤代表一种用于改变所有权组的所有 As described above, the step of FIG. 4 represent a description of a method for changing the ownership of all groups of

权的技术。 Technology rights. 在这种技术中,步骤402需要所有权变化机制一直等待, In this technique, step 402 requires the ownership change mechanism has been waiting for,

直到所有事务都已提交或是回滚,其中这些事务改变了归属于转移中的所有权组的数据之外的那些数据。 Until all transactions have been committed or rolled back, where these transactions attributable to changes in those data outside the data transfer ownership of the group. 在这个等待过程中,转移中 In this waiting process, metastasis

的所有权组的所有数据都是不可用的。 Ownership of all data sets are not available. 因此,最小化等待时间是非 Thus, to minimize non-latency

常重要的。 Very important.

如上所述,对实际改变了归属于转移中的所有权组的数据的事务进行追踪,这很可能是不切实际的。 As mentioned above, the actual change in the ownership belongs to the group of data transferred in a transaction tracking, which is likely to be impractical. 因此,所有权变化机制等待所有运行在归属于转移中的所有冲又组的源所有者集合的所有数据库服务器上的事务提交或者回滚。 Therefore, the change of ownership mechanism waiting for transactions to commit or rollback on all database servers running on all source owners rushed attributable to the transfer of another set of groups. 由于所有4又变化4几制必须等待的事务数量,并且其中许多事务有可能并未改变转移中的所有权组的数据,因此延迟有可能是非常重要的。 Since all changes 4 4 and made a few number of transactions must wait, and there are many matters which may not change the data transfer in the ownership group, so delays may be very important.

根据一个替换实例,提供了一种机制,该机制允许那些在所有者之间转移的数据在这个延迟中保持可用。 According to an alternative example, there is provided a mechanism that allows the owner of that data transfer between the holder may be used in this delay. 具体地说,禁止变化的 In particular, the ban changes

消息并未发送到所有数据库服务器。 Message not sent to all database servers. 相反, 一个"新的所有者"消息发送到所有数据库服务器,它指示的是所有权组的目标所有者集合。 In contrast, a "new owner" message is sent to all database servers, which indicates that the owner of the ownership group of the set target. 新的所有者消息可以广播,举例来说,在更新控制文件之后向所有数据库服务器发送一个刷新緩存的消息,以便指示:(l)源所有者集合,(2)目标所有者集合,以及(3)该所有权组处在转移之中。 The new owner may broadcast a message, for example, send a message to all database refresh the cache server after the update control file to indicate: (l) set the owner of the source, (2) set the owner of the target, and (3 ) the ownership group is in being transferred.

在服务器接收到新的所有者的消息之后,由该服务器启动的所有事务都开始行动,就好像目标所有者集合拥有该所有权组那样。 After the server receives a new message to the owner of all transactions initiated by the server they are beginning to act as if the owner of the collection target of the ownership group that owns.

40在服务器接收新的所有者消息之前,服务器中已经启动的所有事务都会继续进行,就好像源所有者集合拥有该所有权组那样。 40 before the server receives the new owner of a message, the server has started all transactions will proceed as if the owner of the source of the ownership group that owns the collection. 因此, 在等待过程中,转移中的所有权组的所有权在源所有者集合成员与目标所有者集合成员之间得到了有效共享。 Therefore, during the waiting period, ownership transfer ownership of the group in the source owner of the collection has been effectively shared between the members and the members of the target collection owner. 换句话说,转移中的所有权组的数据是在两个数据库服务器之间临时共享的,并且共享磁盘锁定机制是为了访问这类数据而被临时激活的。 In other words, the data transfer ownership group is in temporary shared between two database servers, and shared disk locking mechanism for access to such data is temporarily activated.

在广播新的所有者消息之前开始的源所有者集合中的所有事务都已提交或者回滚的时候,控制文件将会再次更新。 When all transaction set source owner before the news began broadcasting the new owner's have been committed or rolled back, the control file will be updated again. 在第二次更新中,控制文件将被更新,以便指示目标所有者集合即为所有权组的独占所有者,并且该所有权组不再处于转移之中。 In the second update of the control file is updated to indicate that the target is the exclusive owner of a set of ownership of the owner group, and the group no longer in the transfer of ownership.

在上述i兌明中,已经参考本发明的实施例而对其加以描述。 In the i against Ming, it has reference to the embodiment of the present invention to be depicted. 然而^f艮明显,可以对本发明进4亍各种4奮改和变4匕,而不脱离它的实质和范围。 However, significant Gen ^ f, right foot 4 can enter various modifications and variations Fen 4 4 dagger of the present invention without departing from its spirit and scope. 因此,说明书以及附图被看作是说明性的,它们并不具有卩艮制意义。 Accordingly, the specification and figures are to be regarded in an illustrative, they do not have the significance Jie Gen system.

Claims (82)

  1. 1.一种数据库系统,包括: 一个或多个永久存储设备,其上储存有数据库; 运行在多个节点上的多个数据库服务器; 其中所述多个节点中的每一个节点都可以直接访问所述一个或多个永久存储设备; 其中至少一部分所述数据库划分为多个所有权组; 其中所述多个所有权组中的每个所有权组都被指派了一个所有者集合; 其中,至少一个所有权组被指派了一个包括所述多个节点中的多于一个节点的所有者集合; 其中只有那些运行在作为一个所有权组的所有者集合的成员的数据库服务器上的进程才被准许直接访问所述所有权组内部的数据。 CLAIMS 1. A database system, comprising: one or more persistent storage devices, which is stored on the database; multiple database servers running on the plurality of nodes; wherein each of said plurality of nodes in a node can directly access the one or more persistent storage devices; wherein said at least a portion of the database into a plurality of groups of ownership; wherein said plurality of title groups each group is assigned ownership of the owner of a set; wherein at least one of ownership comprising a group owner is assigned more than one node in the set of the plurality of nodes; wherein only those processes that are running on the database server as an owner of a set of ownership group was only permits direct access to the internal data ownership group.
  2. 2. 根据权利要求l所述的数据库系统,其中:所述多个所有权组中的每个所有权组都净皮指定为无共享所有纟又纟且或共享^兹盘的所有纟又组;每个无共享所有权组都#:指派了一个来自所述多个数据库服务器中的所有者;只有每个无共享所有权组的所有者才被准许直接访问所述无共享所有权组内部的凄t据;以及所述多个数据库服务器中的每一个都被准许直接访问指定为共享》兹盘的所有权组的所有4又组内部的凄t据。 The database system according to claim l, wherein: the plurality of title groups each title group are given no net shared by all skin and Si or Si and Si ^ All shared disks and hereby group; each a shared nothing ownership group are #: assigned from the owner of a multiple database servers; not only the owner of the shared ownership of each group was only allowed to directly access the internal desolate no shared ownership group according to t; and the plurality of database servers each have been granted direct access to data designated as shared sad t 4 and all internal group "ownership group hereby disk.
  3. 3. 根据权利要求2所述的数据库系统,还包括一种结构,用于将指定为共享》兹盘所有权组改变成指定为无共享所有权组,其中所述结构包括所述数据库系统中负责改变所述指定的一部分。 3. The database system of claim 2, further comprising a structure for changing the specified shared ownership to a designated group without sharing "hereby disk ownership group, wherein said structure comprises a database system is responsible for the change the designated portion.
  4. 4. 4艮据4又利要求3所述的凝:据库系统,其中所述结构-故配置成响应于确定所述共享磁盘的所有权组包含一个写热点,将指定为共享^f兹盘所有权组自动改变成指定为无共享所有权组。 4.4 Gen condensate according to the claims 4 and 3: database system, wherein the structure - it is configured, in response to determining that the ownership of the shared disk group contains a write hot, will hereby designated as the shared disk ^ f automatic change to the ownership group designated as shared-ownership group.
  5. 5. 根据权利要求2所述的数据库系统,还包括一种结构,用于将指定为无共享所有权组改变成指定为共享磁盘的所有权组。 The database system according to claim 2, further comprising a structure for the group given no change to the specified shared ownership of the shared disk ownership group.
  6. 6. 根据权利要求5所述的数据库系统,其中所述结构被配置成响应于确定所述无共享所有权组包含一个读热点,将指定为无共享所有权组自动改变成指定为共享磁盘的所有权组。 6. A database system according to claim 5, wherein the structure is configured in response to determining that the group contains no shared ownership of a read hot-spot, the group given no shared ownership automatically changed to the specified shared disk ownership group .
  7. 7. 根据权利要求3所述的数据库系统,其中所述结构还4皮配置成将指定为无共享所有权组改变成指定为共享磁盘的所有权组。 7. A database system according to claim 3, wherein said skin structure further 4 configured as a non-designated group change to the specified shared ownership of the shared disk ownership group.
  8. 8. 根据权利要求7所述的数据库系统,其中所述结构被配置成响应于确定所述无共享所有权组包含一个读热点,将指定为无共享所有权组自动改变成指定为共享》兹盘的所有权组。 8. The database system of claim 7, wherein the structure is configured in response to determining that the group contains no shared ownership of a read hot-spot, the group given no shared ownership share is automatically changed to the specified "hereby disc ownership group.
  9. 9. 根据权利要求2所述的数据库系统,还包括一个分布式锁定管理器,其^皮配置成对指定为共享》兹盘所有^又组的所有^又组内部数据访问进行管理,而不对指定为无共享所有权组的所有权组内部数据访问进行管理。 9. The database system of claim 2, further comprising a distributed lock manager, which is configured to transdermal ^ designated as shared "tray all hereby ^ ^ and set all internal data access and manage the group, but not for designated as shared-ownership group of internal data access ownership group management.
  10. 10. —种对一个或多个永久存储设备中保存的数据库的访问进行管理的方法,其中所述永久存储设备能够直接访问多个节点上运行的多个数据库服务器,所述方法包括以下步骤:将至少一部分所述数据库划分为多个所有权组;将一个或多个所有者集合中的至少一个所有者集合指派给所述多个所有权组中的每一个所有权组;以及只允许那些运行在属于每个所有权组的所有者集合的数 10. The - method of access to the database stored in one or more persistent storage devices for management, wherein the permanent storage device capable of directly accessing a plurality of database servers running on multiple nodes, said method comprising the steps of: at least a portion divided into a plurality of database ownership groups; at least one owner or a set of multiple owners set assigned to the plurality of title groups each of a set of ownership; and only run those belonging to the owner of each number set ownership group
  11. 11. 根据权利要求10所述的方法,其中将一个或多个所有者集合中的至少一个所有者集合指派到所述多个所有权组中的每个所有权组的步骤包括:将所述多个所有权组中的每个所有权组指定为无共享所有4又组或共享》兹盘的所有纟又组;为每个无共享所有权组指派仅一个来自所述多个数据库服务器的所有者;只允许每个无共享所有权组的所有者直接访问所述无共享所有权组内部的数据;以及允许属于每个所有权组的所有者集合的所述多个数据库月良务器中的每一个直4矣访问指定为共享》兹盘的所有^又组的所有权组内部的数据。 11. The method of claim 10, wherein one or more of the at least a set of the owner of the owner of the collection assigned to each step group ownership of the ownership of the plurality of groups comprises: the plurality of ownership of each group given no shared ownership group and all 4 groups or sharing "disks and all hereby Si group; no shared ownership for each group is assigned only one owner of the database server from the plurality; only allows each group owner without shared ownership of the direct access to the internal shared data without ownership group; and allowing direct access to each of said plurality of carry 4 months good service database belonging to each set of the owner in the ownership group ^ all specified data inside the group and group ownership of a shared "hereby disk.
  12. 12. 根据权利要求11所述的方法,还包括将指定为共享磁盘的所有权组改变成指定为无共享所有权组的步骤。 12. The method according to claim 11, further comprising a disk designated as the shared ownership to a designated group change group ownership of shared-step.
  13. 13. 根据权利要求12所述的方法,其中响应于确定所述共享磁盘所有权组包含一个写热点,自动执行将指定为共享》兹盘的所有权组改变成指定为无共享所有^又组的步骤。 13. The method according to claim 12, wherein in response to determining that the shared disk ownership group comprising hotspot a write is performed automatically specified as shared "hereby changed to disk ownership group given no ^ and shared by all groups .
  14. 14. 根据权利要求l 1所述的方法,还包括将指定为无共享所有权组改变成指定为共享》兹盘的所有4又组的步骤。 14. l The method of claim 1, further comprising a shared ownership group given no change to step 4 and all groups designated as shared "hereby disc.
  15. 15. 4艮据4又利要求14所述的方法,其中响应于确定所述无共享所有4又组包含一个读热点,自动执4亍爿寻指定为无共享所有4又组改变成指定为共享磁盘的所有权组的步骤。 Gen 4 and 15.4 According to the method of claims 14, wherein in response to determining that the non-shared for all four groups further contains a read hot-spot, the right foot valves 4 is automatically executed to find given no shared by all groups was changed to 4 and designated as sharing the step of disk ownership group.
  16. 16. 根据权利要求12所述的方法,还包括将指定为无共享所有权组改变成指定为共享^兹盘的所有一又组的步骤。 16. The method of claim 12, further comprising a non-designated group change to the specified shared ownership share all the steps of a disc and hereby ^ group.
  17. 17. 根据权利要求16所述的方法,其中响应于对所迷无共孚所有权组包含一个读热点的判定,自动执行将指定为无共享所有权组改变成指定为共享磁盘的所有权组的步骤。 17. The method according to claim 16, wherein the fan in response to the non-proprietary common Fu group comprising hotspot is determined that a read is performed automatically be designated as non-shared ownership rights group is changed to the step of groups designated as shared disk.
  18. 18. 根据权利要求11所述的方法,还包括使用一个分布式锁定管理器来对指定为共享》兹盘的所有权组的所有权组内部数据访问进^f亍管理,而不对指定为无共享所有^K组的所有^UE内部凌丈据访问进行管理的步骤。 18. The method according to claim 11, further comprising the use of a distributed lock manager to be designated as shared "internal group ownership of the ownership of the data set is hereby disk access management right foot into ^ f, not designated as non-shared by all All internal Ling Zhang ^ UE ^ K group of data access step management.
  19. 19. 一种数据库系统,包括:数据库;多个数据库服务器;其中所述多个数据库服务器中的每个数据库服务器都可以直4妻访问所述数据库;其中至少一部分所述数据库#1划分成多个所有^又组;其中所述多个所有权组中的至少一个所有权组被指派了一个所有者集合;以及其中在作为所有权组所有者集合成员的数据库服务器上运行的进程被准许直接访问所述所有权组内部的数据。 19. A database system, comprising: a database; a plurality of database servers; wherein said plurality of database servers each database server can access the database directly wife 4; wherein at least a portion divided into a plurality of database # 1 a ^ and all groups; wherein the plurality of ownership groups in at least one ownership group is assigned a set of owner; and processes which run on the database server as a member of the ownership group of the owner of the collection is granted direct access to the internal data ownership group.
  20. 20. 冲艮据权利要求19所述的数据库系统,其中所述多个所有4又组的每个所有权组都被指派了一个所有者集合。 20. A punch, according to claim Gen database system of claim 19, wherein each of said plurality of all group 4 and group ownership is assigned a set of the owner.
  21. 21. 根据权利要求19所述的数据库系统,其中只有那些运行在作为一个所有权组所有者集合成员的数据库服务器上的进程才被准许直4妄访问所述所有4又组内部的凄t据。 21. The database system according to claim 19, wherein only those processes running on the database server as a member of the ownership group of the owner of the collection was only allowed to direct all sad inside, according to t 4 and Group 4 jump to the access.
  22. 22. 根据权利要求19所述的数据库系统,其中所述多个所有权组的至少一个所有片又症且包4舌一个或多个表空间。 22. A database system according to claim 19, wherein the at least one sheet of the plurality of ownership of all groups of packets 4 and the tongue disease and one or more tablespaces.
  23. 23. 根据权利要求22所述的数据库系统,其中一个或多个表空间中的至少一个表空间是一个数据文件的集合。 23. A database system according to claim 22, wherein the one or more tablespaces at least one table space is a collection of data files.
  24. 24. 根据权利要求19所述的数据库系统,其中:所述多个所有4又组中的至少一个所有4又组;故指定为无共享所有纟又纟且;至少一个无共享所有4又组^皮/人所述多个凄丈据库月l务器中指》泉了一个所有者;以及只有每个无共享所有纟又组的所有者才^皮准i牛直4妻-沐问所述无共享所有权组内部的数据。 24. The database system according to claim 19, wherein: all of said plurality of groups of 4 and 4 and at least one of all the groups; therefore given no Si and Si and shared by all; at least one non-shared and all 4 groups ^ leather / person the more desolate husband databases month l Services is the middle finger, "Quan an owner; the owner and only each shared-Si and all groups did ^ i cow leather quasi-straight 4 wives - ask the Mu no data inside said shared ownership group.
  25. 25. 根据权利要求19所述的数据库系统,其中:所述多个所有权组的至少一个所有权组被指定为共享磁盘的所有一又组;以及所述多个数据库服务器中的每一个都被准许直接访问那些被指定为共享磁盘所有权组的所有权组内部的数据。 25. The database system according to claim 19, wherein: at least one of said plurality of title group ownership of all groups designated and a group of shared disks; and each of said plurality of database servers are permitted direct access to the data inside those ownership group is designated as a shared disk ownership group.
  26. 26. 根据权利要求19所述的数据库系统,其中:所述多个所有权组中的每个所有权组都被指定为无共享所有4又组或是共享》兹盘的所有4又组;以及每个无共享所有权组都从所述多个数据库服务器中指派了一个单独的所有者。 26. The database system according to claim 19, wherein: said plurality of title groups each group is assigned ownership of the shared-free and all 4 groups or sharing "and all 4 groups hereby disc; and each of a non-shared ownership group are assigned a single owner from the plurality of database servers.
  27. 27. 根据权利要求19所述的数据库系统,其中所述多个所有权组中的至少一个所有权组被指定为多种类型的所有权组中的某个特定类型的所有权组。 27. A database system according to claim 19, wherein at least one of said plurality of title groups ownership group is designated for a particular type of ownership of a plurality of types of group ownership group.
  28. 28. 根据权利要求27所述的数据库系统,还包括一种结构,用于改变所指定的所有权组的特定类型,其中所述结构包括所述数据库系统中负责改变所述所有4又组的特定类型的一部分。 28. A database system according to claim 27, further comprising a structure for a particular type of ownership to change the designated group, wherein said structure comprises a database system 4 is responsible for all changes to the specific groups and part type.
  29. 29. 根据权利要求27所述的数据库系统,其中:所述多个所有权组中的至少一个所有权组被指定为所述多种类型的所有权组中的第一类型的所有权组;以及所述多个所有权组中的至少一个所有权组被指定为所述多种类型的所有权组中的第二类型的所有权组。 29. The database system according to claim 27, wherein: at least one of said plurality of title group ownership group is designated as the group of the plurality of types of ownership rights of the first type of groups; and the plurality a group of at least one title set is designated as the ownership of the ownership of the plurality of types of the second type of group ownership group.
  30. 30. 根据权利要求29所述的数据库系统,其中:所述第一类型的所有权组是一个无共享所有权组;以及所述第二类型的所有^又组是一个共享》兹盘的所有纟又组。 30. The database system according to claim 29, wherein: the first type is a non-shared ownership group ownership group; and the second type is a set of all ^ and share all Si "disc and hereby group.
  31. 31. 才艮据^L利要求30所述的^t据库系统,其中所述多个所有;f又组中的每个所有一又组都#皮指定为无共享所有4又组或是共享》兹盘的所有纟又组。 31. It was Gen ^ t ^ L of claims database system of claim 30, wherein all of said plurality; F and each group one and all the groups were given no skin # 4 and shared by all groups or share "All Si hereby disk and groups.
  32. 32. 根据权利要求27所述的数据库系统,其中,对于多种类型的所有权组中的至少一种类型的所有权组来说,在指定为至少一种类型的所有权组的每个所有权组的所有者集合中只提供了多个数据库月良务器中的一个凄史据库服务器。 32. The database system according to claim 27, wherein, for at least one type of group of a plurality of types of ownership rights of the group, all as designated at least one type of ownership of each group of group ownership who set provides a sad history of multiple databases month good service is the only database server.
  33. 33. 根据权利要求27所述的数据库系统,其中,对于多种类型的所有权组中的至少一种类型的所有权组来说,多个数据库服务器种的每个数据库服务器都包括在指定为至少一种类型的所有权组的每个所有权组的所有者集合中。 33. The database system according to claim 27, wherein, for at least one type of group of a plurality of types of ownership rights of the group, the plurality of kinds of database servers each database server comprises at least one designated the owner of each ownership group of the ownership group of types of collection.
  34. 34. 根据权利要求27所述的数据库系统,其中,对于多种类型的所有权组中的至少一种类型的所有权组来说,多个数据库服务器中至少两个且少于全部的彩:据库力艮务器包4舌在指定为至少一种类型的所有权组的每个所有权组的所有者集合中。 34. The database system according to claim 27, wherein, for at least one type of group of a plurality of types of ownership rights of the group, the plurality of database servers and less than all of the at least two color: database pack force that works to set the owner of the tongue 4 of each group ownership of the specified group ownership of at least one type.
  35. 35. 根据权利要求19所述的数据库系统,还包括:多个数据库服务器中的第一数据库服务器,其中所述第一数据库服务器希望得到那些包括在指派给某个所有者集合的某个所有权组中的数据;其中,如果所述第一数据库服务器包括在所述特定的所有者集合中,则在所述第一数据库服务器上执行的进程将会直接检索数据;以及其中,如果所述第一数据库服务器并没有包括在所述特定所有者集合之中,则在所述第一数据库力良务器上执行的进程从所述多个数据库服务器中的第二数据库服务器请求和接收数据,其中所述第二数据库服务器包括在所述特定的所有者集合之中。 35. The database system according to claim 19, further comprising: a plurality of database servers in the first database server, wherein the first database server that are included in the hope that a group is assigned to a title set of the owner the data; wherein, if the first server includes a database in the specific set of the owner, a process is performed on the first database server will retrieve the data directly; and wherein, if the first the database server does not include the process of requesting and receiving data from the second plurality of database servers in the server database owner specific set, is executed on the first good traffic database force, wherein the said second database server included in the specific set of the owner.
  36. 36. 根据权利要求35所述的数据库系统,其中所述特定的所有权组是一个共享磁盘的所有权组,所述特定的所有者集合包括多个数据库服务器,并且在所述第一数据库服务器上执行的进程直接冲企索^:据。 36. A database system according to claim 35, wherein ownership of the particular group is a set of disks shared ownership of the particular owner set includes a plurality of database servers, and executed on the first database server the process of direct red cable companies ^: It is.
  37. 37. 根据权利要求35所述的数据库系统,其中所述特定的所有权组是一个无共享所有权组,并且多个数据库服务器中的第二数据库服务器是所述特定所有者集合中的唯一的数据库服务器。 37. A database system according to claim 35, wherein the particular group is a non-shared ownership of the ownership group, and a plurality of database servers in the second database server is the only owner of a particular set of database servers .
  38. 38. 根据权利要求35所述的数据库系统,其中所述特定的所有权组是一个无共享所有权组,并且多个数据库服务器中的第一数据库服务器是所述特定所有者集合中的唯一的数据库服务器。 38. A database system according to claim 35, wherein the particular group is a non-shared ownership of the ownership group, and a plurality of database servers in the first database server is the only owner of a particular set of database servers .
  39. 39. 根据权利要求19所述的数据库系统,其中多个所有权组中的至少一个所有权组-故指派了来自所述多个数据库服务器之中的一个所有者,并且其中所述凄史才居库系统还包^":一种结构,用于4巴至少一个所有权组的所有者从多个数据库月l务器中的第一^:据库月l务器重新指派到多个^:据库月l 务器中的第二数据库服务器,其中所述结构包括所述数据库系统中负责重新指派所述所有者的一部分。 39. A database system according to claim 19, wherein at least one group of the plurality of ownership of the ownership group - it is assigned an owner from among the plurality of database servers, and wherein only the sad home history database the system further ^ ": a structure for ownership of the owner of a group at least 4 bar, from a plurality of service database May l ^ a first vessel: May l Services database re-assigned to the plurality ^: database l month in the second service is a database server, wherein said structure comprises a portion of said database system, the owner is responsible for reassignment.
  40. 40. 根据权利要求39所述的数据库系统,其中至少一个所有权组是无共享所有纟又组。 40. A database system according to claim 39, wherein at least one group is a non-shared ownership of all groups and Si.
  41. 41. 根据权利要求39所述的数据库系统,其中所述结构被配置成响应于"i青求而为至少一个所有纟又组重新指派所有者。 41. A database system according to claim 39, wherein the structure is configured to respond to "i cyan demand and Si and at least one group reassign all owners.
  42. 42. 根据权利要求39所述的数据库系统,其中所述结构被配置成响应于所述第一^:据库服务器故障而为至少一个所有权组重新指派所有者。 42. A database system according to claim 39, wherein the structure is configured to respond to the first ^: database server failure reassign ownership of at least one group owner.
  43. 43. 根据权利要求19所述的数据库系统,还包括:一种结构,用于将某个所有权组从第一所有者集合转移到第二所有者集合,其中所述结构被配置成:指示所述多个数据库服务器停止创建特定所有权组内部新版本的数据;以及当通过所述第一所有者集合而对所述数据进行访问所有事务都已提交或中断的时候,改变那些表示所述特定所有权组所有权的数据,以便指示所述第二所有者集合即为所述特定所有权组的所有者,其中,所述结构包括所述数据库系统中负责转移所迷某个所有才又组的一部分。 43. A database system according to claim 19, further comprising: a structure for a set of ownership transferred from the first group to the second owners Owner set, wherein the structure is configured to: indicate the the plurality of database server stops to create a new version of the internal ownership of a particular set of data; and when all transactions and access to the data by the first owner of the collection have been submitted or interrupted, indicating the change of ownership of those particular group ownership of the data, so as to indicate the second set of the owner is the owner of the ownership of the particular set, wherein said structure comprises a portion of a turn until all the groups responsible for the transfer of the database system is lost.
  44. 44. 4艮据4又利要求43所述的凄史据库系统,其中响应于所述结构将所述特定所有权组从所述第一所有者集合转移到所述第二所有者集合时出现的故障,所述结构还被配置成:确定^t障是否发生在改变那些指示所述特定所有;K组所有4又的凄t据之前;如果所述故障发生在改变指示所述特定所有权组的所有权的数据之前,则将所述第一所有者集合恢复为所述特定所有4又纟且的所有者;以及如果所述故障发生在改变指示所述特定所有权组的所有权的数据之后,则将所述第二所有者集合保持为所述特定所有权组的所有者。 44.4 appears Burgundy claims 4 and sad history database system of claim 43, wherein in response to the specific structure of the title set is transferred from the first to the second set of the owner set owner fault, the structure is further configured to: determine whether a fault occurred ^ t that indicates that the particular change all; K until all group 4 data and the bitter t; indicating that the specific change set if the ownership fault occurs before the ownership of the data, then the first set of the owner to recover all of the particular owner and 4 and Si; and if after a change of ownership of the data indicative of the particular group to which the ownership fault occurs, the holding said second set owner as an owner of the particular set of ownership.
  45. 45. 根据权利要求19所述的数据库系统,还包括:一种结构,用于将某个所有权组从第一所有者集合转移到第二所有者集合;以及一种查询,所述查询包含在所述^U居库中,其中所述查询与一个涉及所述第一所有者集合的执行计划相关联,并且其中,在通过所述结构而将所述特定的所有权组从所述第一所有者集合移动到所述第二所有者集合之后,将会生成一个涉及所述第二所有者集合的新的执行计划;其中,所述结构包括所述数据库系统中负责转移所述某个所有纟又纟且的一部分。 45. The database system according to claim 19, further comprising: a structure for a set of ownership transferred from the first group to the second owners Owner set; and one query included in the query ^ U ranks in the library, where the query execution plan is associated with a set of related to the first owner, and where, all through the structure and the particular ownership group from the first after moving to the second set of who the owner of the collection, will generate a new execution plan of the second set of owners involved; wherein the structure of the database system, including the one responsible for the transfer of all Si and another part of Si.
  46. 46. 根据权利要求19所述的数据库系统,还包括:一种结构,用于将数据项从第一所有权组重新指派到第二所有4又组,其中所述结构-故配置成:禁止多个数据库服务器访问所述数据项;以及当所有正在访问所述数据项的事务都已提交或者中断的时候,改变指示所述凄t据项所属的所有4又组的^t据,以指示所述据项属于所述第二所有^f又组,其中,所述结构包括所述数据库系统中负责重新指派所述凄t据项的一部分。 46. ​​The database system according to claim 19, further comprising: a structure for a data item from the first reassign ownership to the second group and all 4 groups, wherein the structure - it is configured to: prohibit multiple database servers to access the data item; and when all is accessing the data items have been submitted to the transaction or interrupt, indicating the change, according to all the sad t ^ t 4 and group data item belongs to indicate said data item belonging to the second group and all ^ f, wherein said structure comprises a database system is responsible for re-assigning the desolate portion t of data items.
  47. 47. 根据权利要求46所述的数据库系统,其中所述结构还被配置成:改变第一数据来指示所述数据项所属的所有权组;以及在改变所述第一数据之前,改变第二数据以指示所述数据项所属的所有^K组。 47. A database system according to claim 46, wherein the structure is further configured to: change the first set of data to indicate ownership of the data item belongs; and before the first change in the data, changing the second data All groups ^ K to indicate that the data item belongs.
  48. 48. 根据权利要求47所述的数据库系统,其中所述第一数据位于一个控制文件中,以及所述第二数据位于一个数据字典中。 48. A database system according to claim 47, wherein said first control data in a file, and the second data in a data dictionary.
  49. 49. 根据权利要求47所述的数据库系统,其中响应于所述结构将所述数据项从所述第一所有权组重新指派到所述第二所有权组时出现的故障,所述结构还净皮配置成:确定所述故障是否出现在改变所述第二数据之前;如果所述古文障出现在改变所述第二凄史才居之前,则将所述数据项恢复到所述第一所有权组;以及如果所述故障出现在改变所述第二数据之后,则通过改变所述第一#:据来完成将所述#:据项从所述第一所有^1组重新指派到所述第二所有^又组。 49. A database system according to claim 47, the data item from the first set of ownership reassigned to a failed when the second set in response to the ownership of the structure, said structure further net Paper configured to: determine the change before the second data if the problem; if the classical barrier appear before it ranks second in the sad history of change, then restore the data item to the first ownership group ; and if the failure occurs after changing the second data, by changing the first #: # data to complete the: first of all the data items from ^ 1 reassigned to the second group All but two ^ group.
  50. 50. #4居4又利要求46所述的教:才居库系统,其中所述结构还^皮配置成:确定所述第二所有纟又组是否正在经历所有;^又改变;以及^口果所述第二所有斗又纟且正在经历所有#又改变,则一夸所述凄t才居项标"i己为移动延迟。 50. # 4 ranks and 4 according to claims 46 to teach: only home database system, wherein the structure further configured to transdermal ^: determining whether the second group is experiencing all Si and all; ^ has changed; and ^ if all the second port and the bucket and Si are undergoing changes and # all, then only a boast the sad home item labeled t "i-hexyl mobile delay.
  51. 51. —种用于管理翁:才居库访问的方法,其中所述凄文才居库可以由多个数据库服务器进行直接访问,所述方法包括以下步骤:至少将所述数据库的一部分划分为多个所有4又组;将一个所有者集合指派给所述多个所有权组中的至少一个所有纟又组;以及允许在属于每个所有权组所有者集合的数据库服务器上运行的进程直接访问每个所有权组内部的数据。 51. - manage Weng species: only home database access method, wherein the sad home Wencai library directly accessed by multiple database servers, said method comprising the steps of: dividing at least a portion of said plurality of database All groups and a 4; a plurality of owners of rights assigned to the set of at least one group of all the groups and Si; and allowing belonging to a process running on database server owner group for each title set direct access to each internal data ownership group.
  52. 52. 4艮据4又利要求51所述的方法,还包4舌以下步骤:将所有者集合指派给所述多个所有权组中的每个所有权组。 Gen 4 and 52.4 According to the method of claims 51, further tongue 4 steps: the ownership of the owner of the group assigned to each set of said plurality of title groups.
  53. 53. 4艮据4又利要求51所述的方法,还包4舌以下步骤:只允许那些在作为某个特定所有权组所有者集合成员的数据库服务器上运行的进程直接访问所述特定所有权组内部的数据。 Gen 4 and 53.4 According to the method of claims 51, further tongue 4 steps: allows only those processes run on the database server as a group owner of a particular title set direct access to members of the particular group ownership internal data.
  54. 54. 根据权利要求51所述的方法,其中所述多个所有权组中的至少一个所有才又组包4舌一个或多个表空间。 54. The method according to claim 51, wherein said plurality of title groups, at least before all of a packet group and a tongue 4 or more tablespaces.
  55. 55. 根据权利要求54所述的方法,其中一个或多个表空间之中的至少一个表空间是一个凝:据文件的集合。 55. The method of claim 54, wherein the one or more among the table space is a space of at least one condensate table: According to a collection of files.
  56. 56. 根据权利要求51所述的方法,还包括以下步骤:将所述多个所有权组中的至少一个所有权组指定为无共享所有4又纟且;向至少一个无共享所有^:又组指派一个来自所述多个凌t才居库服务器之中的所有者;以及只允i午每个无共享所有4又组的所有者直4妻访问所述无共享所有相j且内部的凝:才居。 56. The method of claim 51, further comprising the steps of: specifying the plurality of title groups is at least one non-shared ownership group and all 4 and Si; at least a share of all the non ^: and a group assignment Ling from a plurality of said owners home t only among library server; i and allow only the owner of all shared-L 4 each group and the non-linear access to shared 4 wife with all j and the interior of the condensate: She was living.
  57. 57. 根据权利要求51所述的方法,还包括以下步骤:将所述多个所有权组中的至少一个所有权组指定为共享磁盘的所有权组;以及允许所述多个数据库服务器中的每一个直接访问那些指定为共享;兹盘的所有4又组的所有— 又组内部的凄史才居。 57. The method of claim 51, further comprising the steps of: specifying the plurality of title groups is at least one group ownership of the shared disk ownership group; and allowing the plurality of database servers each of a direct those designated as shared access; all in all hereby disk 4 and the group - and the sad history of the group was inside the residence.
  58. 58. 根据权利要求51所述的方法,还包括以下步骤:将所述多个所有权组中的每一个所有权组指定为无共享所有4又组或共享,兹盘的所有斥又组;以及从所述多个数据库服务器中为每个无共享所有权组指派一个单独的所有者。 58. The method of claim 51, further comprising the steps of: said plurality of title groups each of a group given no ownership share and all 4 groups or shared, all hereby disc and repellent groups; and from the plurality of database servers is assigned a single owner for each non-shared ownership group.
  59. 59. 根据权利要求51所述的方法,还包括以下步骤:将所述多个所有权组中的至少一个所有权组指定为多种类型的所有;MUa中的某种特定类型的所有^=又组。 59. The method of claim 51, further comprising the steps of: specifying the at least one group of the plurality of ownership of the ownership group is all of a plurality of types; MUa in a specific type of group and all ^ = .
  60. 60. 4艮据权利要求59所述的方法,还包括以下步骤:改变所指定的所有权组的类型。 The method of claim 59 60.4 Gen according to claim, further comprising the step of: changing the type of ownership of the specified group.
  61. 61. 根据权利要求59所述的方法,还包括以下步骤:将所述多个所有权组中的至少一个所有权组指定为所述多种类型的所有权组中的第一类型的所有权组;以及将所述多个所有权组中的至少一个所有权组指定为所述多种类型的所有权组中的第二类型的所有权组。 61. The method according to claim 59, further comprising the steps of: specifying the at least one group of the plurality of ownership of the ownership of the ownership group is the group of the first type of the plurality of types of ownership of the group; and at least one title group as the group of the plurality of types of ownership of the ownership of a second set of said plurality of types of ownership groups.
  62. 62. 根据权利要求61所述的方法,其中:所述第一类型的所有权组是无共享所有权组;以及所述第二类型的所有4又组是共享》兹盘的所有4又组。 62. The method according to claim 61, wherein: the first type groups are not shared ownership rights group; and the second type and all 4 groups and 4 is shared by all groups "hereby disc.
  63. 63. 根据权利要求62所述的方法,还包括以下步骤:将所述多个所有权组中的每个所有权组都指定为无共享所有片又组或是共享》兹盘的所有纟又组。 63. The method according to claim 62, further comprising the steps of: said plurality of title groups each of the groups were given no ownership share all pieces and share all or groups Si "and hereby disk group.
  64. 64. 根据权利要求59所述的方法,还包括以下步骤:对于所述多种类型的所有权组的至少一种类型的所有权组来说,仅仅为指定为至少一种类型的所有权组的每个所有权组提供了处于所有者集合的多个数据库服务器之中的一个数据库服务器。 64. The method of claim 59, further comprising the step of: for said plurality of types of ownership of the group at least one type of group ownership, the only designated for each group of at least one type of ownership ownership among the group provides multiple database server is the owner of the collection of a database server.
  65. 65. 根据权利要求59所述的方法,还包括以下步骤:对于所述多种类型的所有权组的至少一种类型的所有权组来说,将多个数据库服务器中的每个数据库服务器都包括在指定为至少一种类型的所有权组的每个所有权组的所有者集合之中。 65. The method according to claim 59, further comprising the step of: for said plurality of types of the at least one type of group for group ownership of the ownership of the plurality of database servers each database server comprises each set being designated as the owner of the ownership group ownership of at least one type of group.
  66. 66. 根据权利要求59所述的方法,还包括以下步骤:对于所述多种类型的所有;K组的至少一种类型的所有斗又组来说,将所述多个数据库服务器中至少两个数据库服务器而少于全部的数据库服务器包括在了标记指定为至少一种类型的所有权组的每个所有权组的所有者集合之中。 66. The method according to claim 59, further comprising the steps of: for all of said plurality of types; all of at least one type of bucket group and K group, the plurality of the at least two database servers database servers and less than all of the database server is included in the collection owner tag specified for each ownership group ownership group of at least one type of being.
  67. 67. 根据权利要求51所述的方法,其中:多个数据库服务器中的第一数据库服务器希望得到包括在指派给某个特定所有者集合的某个特定所有权组的数据;如果所述第一凄t据库力l务器包括在所述特定的所有者集合中,则运行在所述第一数据库服务器上的进程将会直接检索所述凄t据;以及如果所述第一凄史据库力艮务器并不包4舌在所述特定的所有者集合中,则运行在所述第一数据库服务器上的进程将会从所述多个数据库服务器中的第二数据库服务器请求和接收数据, 其中所述第二数据库服务器包括在所述特定的所有者集合中。 67. The method according to claim 51, wherein: the plurality of first database server comprises a database server wishes to obtain ownership of a particular data set is assigned to a particular set of the owner; if the first sad t l Services database force comprises a process in the specific set of the owner, the first run on the database server will directly retrieving the data sad t; and if the first history database sad force that works to the tongue 4 does not process the packet in a particular owner of the collection, the first run on the database server, database server will request and receive second data from the plurality of database servers wherein said second server comprises a database in the specific set of the owner.
  68. 68. 根据权利要求67所述的方法,其中所述特定的所有权组是一个共享磁盘的所有权组,所述特定的所有者集合包括所述多个数据库服务器,并且运行在所述第一数据库服务器上的进程直接检索所述数据。 68. The method of claim 67, wherein ownership of the particular group is a set of disks shared ownership of the particular owner of the set comprises a plurality of database servers, and running the first database server process on the data retrieved directly.
  69. 69. 根据权利要求67所述的方法,其中所述特定所有权组是一个无共享所有权组,所述多个数据库服务器中的第二数据库服务器是所述特定的所有者集合中的唯一的数据库服务器。 69. The method according to claim 67, wherein the particular group is a non-shared ownership of the ownership group, said second plurality of database servers in the particular database server is the only owner of a set of database servers .
  70. 70. 才艮据权利要求67所述的方法,其中所述特定所有外又组是一个无共享所有权组,所述多个数据库服务器中的第一数据库服务器是所述特定所有权组中的唯一的数据库服务器。 70. The method according to claim Burgundy was claimed in claim 67, wherein said outer and specific groups are all non-shared ownership of a group, the plurality of database servers in the first database server is the particular set of unique ownership the database server.
  71. 71. 才艮才居4又利要求51所述的方法,还包4舌以下步骤:从所述多个数据库服务器中向所述多个所有权组中的至少一个所有一又纟且指》泉一个所有者;以及将所述至少一个所有权组的所有者从所述多个数据库服务器的第一数据库服务器重新指派到所述多个数据库服务器中的第二数据库服务器。 Burgundy was only 71. 4 ranks and claims method of claim 51, further tongue 4 steps: from the plurality of database servers and at least one of Si, and all of a plurality of ownership refers to the groups "spring a owner; and the at least one group owner re-assign ownership database server from a first server of the plurality of database server database to a second database of the plurality of servers.
  72. 72. 根据权利要求71所述的方法,其中所述至少一个所有权组是无共享所有权组。 72. The method according to claim 71, wherein said at least one group is a non-shared ownership rights group.
  73. 73. 根据权利要求71所述的方法,其中为所述至少一个所有权组重新指派所有者集合的步骤是响应于请求而被执行的。 73. The method according to claim 71, wherein the at least one step of reassigning ownership of the owner of the group is set to be executed in response to the request.
  74. 74. 才艮据;〖又利要求71所述的方法,其中为所述至少一个所有冲又组重新指派所有者集合的步骤是响应于所述第一数据库服务器故障而#皮自动#^亍的。 74. It was Gen; 〖method of claim 71 further claims, wherein the at least one punch and all groups re-assignment step is set owner and automatically # # skin in response to the first database server failure ^ right foot of.
  75. 75. 根据权利要求51所述的方法,还包括以下步骤:通过执行以下步骤而将某个特定所有权组从第一所有者集合转移到第二所有者集合:命令所述多个数据库服务器停止创建所述特定所有权组内部新版本的凄t据;以及当所有经由所述第一所有者集合来访问所述数据的事务都已提交或是中断的时候,改变那些指示所述特定所有权组的所有权的数据,以便指示所述第二所有权组是所述特定所有权组的所有者。 75. The method according to claim 51, further comprising the step of: while the ownership of a specific group set by performing the steps of transferring from the first to the second owners Owner set of: instructing the plurality of database servers stop creating the new version of the interior of the particular ownership group according to t desolate; and when all the time through the first set of owners to access the data of the transaction have been submitted or interrupted, indicating the change of ownership of those particular ownership group data, to indicate the owner of the second set of ownership rights of the particular group.
  76. 76. 4艮据权利要求75所述的方法,还包括以下步骤:响应于将所述特定所有权组从所述第一所有者集合转移到所述第二所有者集合时出现的故障,确定所述故障是否是在改变了指示所述特定所有权组所有权的数据之前出现的;如果所述故障出现在改变指示所述特定所有权组所有4又的凄t据之前,则将所述第一所有斥又组恢复为所述特定所有斗又组的所有者;以及如果所述故障出现在改变指示所述特定所有权组所有权的数据之后,则将所述第二所有者集合保持为所述特定所有权組的所有者。 76.4 The method according to claim Gen 75, further comprising the step of: in response to the particular ownership transferred from the first group set to a failure when the owner of the second set of the owner, to determine the if said fault occurs before the change is indicative of the specific title data set of ownership; if the fault occurs in the ownership of the particular set change instruction and all the bitter t 4 before the data, then all the first-repellent and return to the particular group and all groups fighting owner; and if the failure occurs after a change in ownership of the particular set of data indicates ownership, then the owner of the second set of retaining ownership of the particular group owner. 16 16
  77. 77. 根据权利要求51所述的方法,其中在所述数据库中包括一个查询,其中所述查询与一个涉及第一所有者集合的执行计划相关写关,并且其中所述方法还包括以下步骤:将某个特定所有权组从所述第一所有者集合转移到第二所有者集合;以及在将所述所有权组从所述第一所有者集合转移到所述第二所有者集合之后,产生一个涉及所述第二所有者集合的新的执行计划。 77. The method according to claim 51, wherein said database comprises in a query, wherein the query execution plan involves a first set of related owners write off, and wherein the method further comprises the steps of: the ownership of a specific group set transferred from the first to the second owners owner set; and after the transfer of ownership from the first group set to the owner of the second set of the owner, to produce a the new implementation plan involving the owner of the second set.
  78. 78. 才艮据斗又利要求51所述的方法,还包括以下步骤:通过执行以下步骤而将一个数据项从第一所有权组重新指派到第二所有权组:禁止所述多个数据库服务器访问所述数据项;以及当访问所述数据项的所有事务都已提交或中断的时候,改变指示所述H悟项所属所有权组的凄t据,以〗更指示所述数据项属于所述第二所有权组。 78. It was Gen bucket and Lee method of claim 51, further comprising the step of: while a data item from the first reassign ownership rights to the second group set by performing the following steps: inhibiting the access to multiple database servers the data item; and when all transactions accessing the data items have been submitted or interrupted, change indicates that the item belongs desolate H Wu ownership group t According to〗 more indicative of the data items that belong to the first two ownership groups.
  79. 79. 根据权利要求78所述的方法,还包括以下步骤:改变第一数据,以便指示所述数据项所属的所有权组;以及在改变第一数据之前,改变第二数据,以便指示所迷数据项所属的所有4又组。 79. The method according to claim 78, further comprising the step of: changing the first data set to indicate ownership of the data item belongs; and prior to the change of the first data, second data change, so that the data indicative of the fan 4 and item belongs to all groups.
  80. 80. 根据权利要求79所述的方法,其中所述第一数据处于一个控制文件之中,所述第二数据处于一个数据字典之中。 80. The method according to claim 79, wherein said first control data is in a file, in said second data into a data dictionary.
  81. 81. 根据权利要求79所述的方法,还包括以下步骤:响应于将所述数据项从所述第一所有权组重新指派到所述第二所有权组时出现的故障,确定所述故障是否是在改变所述第二数据之前发生的;如果所述故障在改变所述第二数据之前发生,则将所述凄t据项恢复到所述第一所有纟又组;以及如果所述故障发生在改变所述第二数据之后,则通过改变所述第一数据来完成所述数据项从所述第一所有权组到所述第二所有权组的重新指派。 81. The method according to claim 79, further comprising the step of: in response to re-assign the data item from the first group to the ownership fault occurs when the ownership of the second set, determining whether the fault is before changing to the second data if the failure occurs, then the sad t data items restored to the first group and all the Si;; occurs before the change of the data, and if the second failure occurs after changing the second data by changing the data to complete the first set of data items from the first to reassign ownership of the ownership of the second group.
  82. 82. 根据权利要求78所述的方法,还包括以下步骤:确定所述第二所有纟又纟且是否正在经历所有4又改变;以及如果所述第二所有权组正在经历所有权改变,则将所述数据项标记为移动延迟。 82. The method according to claim 78, further comprising the step of: determining that all of the Si and second Si and is experiencing all changes and 4; and if the second group is experiencing a change of ownership rights, then the said delay data item is marked as moved.
CN 01822844 1998-02-12 2001-06-28 Partitioning ownership of a database among different database servers to control access to the database CN100517303C (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/US2001/020842 WO2003003252A1 (en) 1998-02-12 2001-06-28 Partitioning ownership of a database among different database servers to control access to the database

Publications (2)

Publication Number Publication Date
CN1505788A true CN1505788A (en) 2004-06-16
CN100517303C true CN100517303C (en) 2009-07-22

Family

ID=28038554

Family Applications (1)

Application Number Title Priority Date Filing Date
CN 01822844 CN100517303C (en) 1998-02-12 2001-06-28 Partitioning ownership of a database among different database servers to control access to the database

Country Status (4)

Country Link
EP (3) EP1521184A3 (en)
JP (1) JP4746838B2 (en)
CN (1) CN100517303C (en)
CA (1) CA2435388C (en)

Families Citing this family (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6850941B1 (en) 2000-12-21 2005-02-01 Vignette Corporation Method and system for native-byte form handling
US7277897B2 (en) 2003-08-01 2007-10-02 Oracle International Corporation Dynamic reassignment of data ownership
CN100449539C (en) 2003-08-01 2009-01-07 甲骨文国际公司 Ownership reassignment in a shared-nothing database system
US7120651B2 (en) 2003-08-01 2006-10-10 Oracle International Corporation Maintaining a shared cache that has partitions allocated among multiple nodes and a data-to-partition mapping
US7139772B2 (en) * 2003-08-01 2006-11-21 Oracle International Corporation Ownership reassignment in a shared-nothing database system
US8234517B2 (en) * 2003-08-01 2012-07-31 Oracle International Corporation Parallel recovery by non-failed nodes
US7814065B2 (en) 2005-08-16 2010-10-12 Oracle International Corporation Affinity-based recovery/failover in a cluster environment
US8510334B2 (en) 2009-11-05 2013-08-13 Oracle International Corporation Lock manager on disk
US8375056B2 (en) 2010-02-26 2013-02-12 International Business Machines Corporation Optimizing data cache when applying user-based security
EP2633462A1 (en) * 2010-10-27 2013-09-04 Max-Planck-Gesellschaft zur Förderung der Wissenschaften e.V. Protecting data integrity with storage leases
JP2012100266A (en) * 2010-10-29 2012-05-24 Kotatsu Kokusai Denshi Kofun Yugenkoshi Method of executing service group owner transfer procedure in communication system
EP2695051A4 (en) * 2011-04-08 2014-11-19 Andrew Liebman Systems, computer readable storage media, and computer implemented methods for project sharing
US20140032595A1 (en) * 2012-07-25 2014-01-30 Netapp, Inc. Contention-free multi-path data access in distributed compute systems
KR101451280B1 (en) * 2012-10-16 2014-10-16 케이티하이텔 주식회사 Distributed database management system and method
US9906410B2 (en) 2014-09-29 2018-02-27 Netapp, Inc. Dynamic data access configuration

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1145489A (en) 1995-06-06 1997-03-19 美国电报电话公司 System and method for database access administration
US5625811A (en) 1994-10-31 1997-04-29 International Business Machines Corporation Method and system for database load balancing

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DE69227665D1 (en) * 1991-10-31 1999-01-07 Ibm A method for operating a computer in a network
DE69619531T2 (en) * 1995-07-21 2002-10-31 Emc Corp Dynamic load balancing
US5907849A (en) * 1997-05-29 1999-05-25 International Business Machines Corporation Method and system for recovery in a partitioned shared nothing database system using virtual share disks
US6594698B1 (en) * 1998-09-25 2003-07-15 Ncr Corporation Protocol for dynamic binding of shared resources

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5625811A (en) 1994-10-31 1997-04-29 International Business Machines Corporation Method and system for database load balancing
US5675791A (en) 1994-10-31 1997-10-07 International Business Machines Corporation Method and system for database load balancing
CN1145489A (en) 1995-06-06 1997-03-19 美国电报电话公司 System and method for database access administration

Also Published As

Publication number Publication date Type
EP2562662A2 (en) 2013-02-27 application
CA2435388A1 (en) 2003-01-09 application
JP2004531005A (en) 2004-10-07 application
CA2435388C (en) 2008-12-09 grant
EP1521184A3 (en) 2006-02-22 application
EP2562662A3 (en) 2015-03-18 application
EP1399847A4 (en) 2005-11-02 application
CN1505788A (en) 2004-06-16 application
EP1521184A2 (en) 2005-04-06 application
JP4746838B2 (en) 2011-08-10 grant
EP1399847A1 (en) 2004-03-24 application
EP1399847B1 (en) 2013-01-16 grant

Similar Documents

Publication Publication Date Title
Rothnie Jr et al. Introduction to a system for distributed databases (SDD-1)
Biliris et al. ASSET: A system for supporting extended transactions
US5515499A (en) Method and system for reconfiguring a storage structure within a structure processing facility
US5287473A (en) Non-blocking serialization for removing data from a shared cache
US6349310B1 (en) Database management system and method for accessing rows in a partitioned table
US6145094A (en) Transaction locks for high availability
US6012059A (en) Method and apparatus for replicated transaction consistency
US6622152B1 (en) Remote log based replication solution
US8103937B1 (en) Cas command network replication
US6170045B1 (en) Cross-system data piping using an external shared memory
US7162467B2 (en) Systems and methods for managing distributed database resources
US6665814B2 (en) Method and apparatus for providing serialization support for a computer system
US20030233370A1 (en) Maintaining a relationship between two different items of data
US6243702B1 (en) Method and apparatus for propagating commit times between a plurality of database servers
US5625811A (en) Method and system for database load balancing
US6976022B2 (en) Method and mechanism for batch processing transaction logging records
US20050160315A1 (en) Geographically distributed clusters
US6219675B1 (en) Distribution of a centralized database
US20040098425A1 (en) Database System Providing Improved Methods For Data Replication
US6981004B2 (en) Method and mechanism for implementing in-memory transaction logging records
US20060136516A1 (en) Techniques for maintaining consistency for different requestors of files in a database management system
US5317739A (en) Method and apparatus for coupling data processing systems
US5642501A (en) Computer method and apparatus for asynchronous ordered operations
US5745753A (en) Remote duplicate database facility with database replication support for online DDL operations
US20040158549A1 (en) Method and apparatus for online transaction processing

Legal Events

Date Code Title Description
C06 Publication
C10 Request of examination as to substance
C14 Granted